Fix CI

Allow strict document update without creating missing documents

.github/ISSUE_TEMPLATE/new_feature_issue.md

@@ -24,6 +24,11 @@ TBD
 - [ ] If not, add the `no db change` label to your PR, and you're good to merge.
 - [ ] If yes, add the `db change` label to your PR. You'll receive a message explaining you what to do.
 
+### Reminders when adding features
+
+- [ ] Write unit tests using insta
+- [ ] Write declarative integration tests in [workloads/tests](https://github.com/meilisearch/meilisearch/tree/main/workloads/test). Specify the routes to call and then call `cargo xtask test workloads/tests/YOUR_TEST.json --update-responses` so that responses are automatically filled.
+
 ### Reminders when modifying the API
 
 - [ ] Update the openAPI file with utoipa:

.github/workflows/bench-manual.yml

@@ -18,7 +18,7 @@ jobs:
     timeout-minutes: 180 # 3h
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/bench-pr.yml

@@ -66,9 +66,7 @@ jobs:
           fetch-depth: 0 # fetch full history to be able to get main commit sha
           ref: ${{ steps.comment-branch.outputs.head_ref }}
 
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
-        with:
-          profile: minimal
 
       - name: Run benchmarks on PR ${{ github.event.issue.id }}
         run: |

.github/workflows/bench-push-indexing.yml

@@ -12,9 +12,7 @@ jobs:
     timeout-minutes: 180 # 3h
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
-        with:
-          profile: minimal
 
       # Run benchmarks
       - name: Run benchmarks - Dataset ${BENCH_NAME} - Branch main - Commit ${{ github.sha }}

.github/workflows/benchmarks-manual.yml

@@ -18,7 +18,7 @@ jobs:
     timeout-minutes: 4320 # 72h
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/benchmarks-pr.yml

@@ -44,7 +44,7 @@ jobs:
             exit 1
           fi
 
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/benchmarks-push-indexing.yml

@@ -16,7 +16,7 @@ jobs:
     timeout-minutes: 4320 # 72h
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/benchmarks-push-search-geo.yml

@@ -15,7 +15,7 @@ jobs:
     runs-on: benchmarks
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/benchmarks-push-search-songs.yml

@@ -15,7 +15,7 @@ jobs:
     runs-on: benchmarks
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/benchmarks-push-search-wiki.yml

@@ -15,7 +15,7 @@ jobs:
     runs-on: benchmarks
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
           profile: minimal
 

.github/workflows/db-change-comments.yml

@@ -19,6 +19,7 @@ env:
 
     - [ ] Detail the change to the DB format and why they are forward compatible
     - [ ] Forward-compatibility: A database created before this PR and using the features touched by this PR was able to be opened by a Meilisearch produced by the code of this PR.
+    - [ ] Declarative test: add a [declarative test containing a dumpless upgrade](https://github.com/meilisearch/meilisearch/blob/main/TESTING.md#typical-usage)
 
 
     ## This PR makes breaking changes
@@ -35,8 +36,7 @@ env:
       - [ ] Write the code to go from the old database to the new one
         - If the change happened in milli, the upgrade function should be written and called [here](https://github.com/meilisearch/meilisearch/blob/3fd86e8d76d7d468b0095d679adb09211ca3b6c0/crates/milli/src/update/upgrade/mod.rs#L24-L47)
         - If the change happened in the index-scheduler, we've never done it yet, but the right place to do it should be [here](https://github.com/meilisearch/meilisearch/blob/3fd86e8d76d7d468b0095d679adb09211ca3b6c0/crates/index-scheduler/src/scheduler/process_upgrade/mod.rs#L13)
-      - [ ] Write an integration test [here](https://github.com/meilisearch/meilisearch/blob/main/crates/meilisearch/tests/upgrade/mod.rs) ensuring you can read the old database, upgrade to the new database, and read the new database as expected
+      - [ ] Declarative test: add a [declarative test containing a dumpless upgrade](https://github.com/meilisearch/meilisearch/blob/main/TESTING.md#typical-usage)
-    
 
 jobs:
   add-comment:

.github/workflows/flaky-tests.yml

@@ -3,7 +3,7 @@ name: Look for flaky tests
 on:
   workflow_dispatch:
   schedule:
-    - cron: '0 4 * * *' # Every day at 4:00AM
+    - cron: "0 4 * * *" # Every day at 4:00AM
 
 jobs:
   flaky:
@@ -13,11 +13,17 @@ jobs:
       image: ubuntu:22.04
     steps:
       - uses: actions/checkout@v5
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
       - name: Install needed dependencies
         run: |
           apt-get update && apt-get install -y curl
           apt-get install build-essential -y
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Install cargo-flaky
         run: cargo install cargo-flaky
       - name: Run cargo flaky in the dumps

.github/workflows/fuzzer-indexing.yml

@@ -12,9 +12,7 @@ jobs:
     timeout-minutes: 4320 # 72h
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
-        with:
-          profile: minimal
 
       # Run benchmarks
       - name: Run the fuzzer

.github/workflows/publish-apt-brew-pkg.yml

@@ -25,7 +25,13 @@ jobs:
         run: |
           apt-get update && apt-get install -y curl
           apt-get install build-essential -y
-      - uses: dtolnay/rust-toolchain@1.89
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Install cargo-deb
         run: cargo install cargo-deb
       - uses: actions/checkout@v5

.github/workflows/publish-docker-images.yml

@@ -208,8 +208,8 @@ jobs:
           done
           cosign sign --yes ${images}
 
-      # /!\ Don't touch this without checking with Cloud team
+      # /!\ Don't touch this without checking with engineers working on the Cloud code base on #discussion-engineering Slack channel
-      - name: Send CI information to Cloud team
+      - name: Notify meilisearch-cloud 
         # Do not send if nightly build (i.e. 'schedule' or 'workflow_dispatch' event)
         if: ${{ (github.event_name == 'push') && (matrix.edition == 'enterprise') }}
         uses: peter-evans/repository-dispatch@v3
@@ -218,3 +218,14 @@ jobs:
           repository: meilisearch/meilisearch-cloud
           event-type: cloud-docker-build
           client-payload: '{ "meilisearch_version": "${{ github.ref_name }}", "stable": "${{ steps.check-tag-format.outputs.stable }}" }'
+
+      # /!\ Don't touch this without checking with integration team members on #discussion-integrations Slack channel 
+      - name: Notify meilisearch-kubernetes
+        # Do not send if nightly build (i.e. 'schedule' or 'workflow_dispatch' event), or if not stable
+        if: ${{ github.event_name == 'push' && matrix.edition == 'community' && steps.check-tag-format.outputs.stable == 'true' }}
+        uses: peter-evans/repository-dispatch@v3
+        with:
+          token: ${{ secrets.MEILI_BOT_GH_PAT }}
+          repository: meilisearch/meilisearch-kubernetes
+          event-type: meilisearch-release
+          client-payload: '{ "version": "${{ github.ref_name }}" }'

.github/workflows/publish-release-assets.yml

@@ -76,7 +76,7 @@ jobs:
     needs: check-version
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Build
         run: cargo build --release --locked ${{ matrix.feature-flag }} ${{ matrix.extra-args }}
       # No need to upload binaries for dry run (cron or workflow_dispatch)

.github/workflows/sdks-tests.yml

@@ -25,14 +25,18 @@ jobs:
       - uses: actions/checkout@v5
       - name: Define the Docker image we need to use
         id: define-image
+        env:
+          EVENT_NAME: ${{ github.event_name }}
+          DOCKER_IMAGE_INPUT: ${{ github.event.inputs.docker_image }}
         run: |
-          event=${{ github.event_name }}
           echo "docker-image=nightly" >> $GITHUB_OUTPUT
-          if [[ $event == 'workflow_dispatch' ]]; then
+          if [[ "$EVENT_NAME" == 'workflow_dispatch' ]]; then
-            echo "docker-image=${{ github.event.inputs.docker_image }}" >> $GITHUB_OUTPUT
+            echo "docker-image=$DOCKER_IMAGE_INPUT" >> $GITHUB_OUTPUT
           fi
       - name: Docker image is ${{ steps.define-image.outputs.docker-image }}
-        run: echo "Docker image is ${{ steps.define-image.outputs.docker-image }}"
+        env:
+          DOCKER_IMAGE: ${{ steps.define-image.outputs.docker-image }}
+        run: echo "Docker image is $DOCKER_IMAGE"
 
 ##########
 ## SDKs ##
@@ -159,24 +163,26 @@ jobs:
       - uses: actions/checkout@v5
         with:
           repository: meilisearch/meilisearch-js
+      - name: Install pnpm
+        uses: pnpm/action-setup@v4
       - name: Setup node
-        uses: actions/setup-node@v5
+        uses: actions/setup-node@v6
         with:
-          cache: 'yarn'
+          cache: 'pnpm'
       - name: Install dependencies
-        run: yarn --dev
+        run: pnpm install
       - name: Run tests
-        run: yarn test
+        run: pnpm test
       - name: Build project
-        run: yarn build
+        run: pnpm build
       - name: Run ESM env
-        run: yarn test:env:esm
+        run: pnpm test:env:esm
       - name: Run Node.js env
-        run: yarn test:env:nodejs
+        run: pnpm test:env:nodejs
       - name: Run node typescript env
-        run: yarn test:env:node-ts
+        run: pnpm test:env:node-ts
       - name: Run Browser env
-        run: yarn test:env:browser
+        run: pnpm test:env:browser
 
   meilisearch-php-tests:
     needs: define-docker-image

.github/workflows/test-suite.yml

@@ -19,87 +19,120 @@ jobs:
     runs-on: ${{ matrix.runner }}
     strategy:
       matrix:
-        runner: [ubuntu-24.04, ubuntu-24.04-arm]
+        runner: [ubuntu-22.04, ubuntu-22.04-arm]
         features: ["", "--features enterprise"]
-    container:
-      # Use ubuntu-22.04 to compile with glibc 2.35
-      image: ubuntu:22.04
     steps:
       - uses: actions/checkout@v5
-      - name: Install needed dependencies
+      - name: check free space before
+        run: df -h
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
         run: |
-          apt-get update && apt-get install -y curl
+          sudo rm -rf "/opt/ghc" || true
-          apt-get install build-essential -y
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - name: check free space after
+        run: df -h
       - name: Setup test with Rust stable
-        uses: dtolnay/rust-toolchain@1.89
+        uses: dtolnay/rust-toolchain@1.91.1
       - name: Cache dependencies
         uses: Swatinem/rust-cache@v2.8.0
-      - name: Run cargo check without any default features
+        with:
+          key: ${{ matrix.features }}
+      - name: Run cargo build without any default features
         uses: actions-rs/cargo@v1
         with:
           command: build
-          args: --locked --release --no-default-features --all
+          args: --locked --no-default-features --all
       - name: Run cargo test
         uses: actions-rs/cargo@v1
         with:
           command: test
-          args: --locked --release --all ${{ matrix.features }}
+          args: --locked --all ${{ matrix.features }}
 
-  test-others:
+  test-windows:
     name: Tests on ${{ matrix.os }}
     runs-on: ${{ matrix.os }}
     strategy:
       fail-fast: false
       matrix:
-        os: [macos-14, windows-2022]
+        os: [windows-2022]
         features: ["", "--features enterprise"]
+    if: github.event_name != 'merge_group'
     steps:
       - uses: actions/checkout@v5
       - name: Cache dependencies
         uses: Swatinem/rust-cache@v2.8.0
-      - uses: dtolnay/rust-toolchain@1.89
+      - uses: dtolnay/rust-toolchain@1.91.1
-      - name: Run cargo check without any default features
+      - name: Run cargo build without any default features
         uses: actions-rs/cargo@v1
         with:
           command: build
-          args: --locked --release --no-default-features --all
+          args: --locked --no-default-features --all
       - name: Run cargo test
         uses: actions-rs/cargo@v1
         with:
           command: test
-          args: --locked --release --all ${{ matrix.features }}
+          args: --locked --all ${{ matrix.features }}
 
-  test-all-features:
+  test-macos:
-    name: Tests almost all features
+    name: Tests on ${{ matrix.os }}
-    runs-on: ubuntu-latest
+    runs-on: ${{ matrix.os }}
-    container:
+    strategy:
-      # Use ubuntu-22.04 to compile with glibc 2.35
+      fail-fast: false
-      image: ubuntu:22.04
+      matrix:
+        os: [macos-14]
+        features: ["", "--features enterprise"]
     if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
     steps:
       - uses: actions/checkout@v5
-      - name: Install needed dependencies
+      - name: Cache dependencies
+        uses: Swatinem/rust-cache@v2.8.0
+      - uses: dtolnay/rust-toolchain@1.91.1
+      - name: Run cargo build without any default features
+        uses: actions-rs/cargo@v1
+        with:
+          command: build
+          args: --locked --no-default-features --all
+      - name: Run cargo test
+        uses: actions-rs/cargo@v1
+        with:
+          command: test
+          args: --locked --all ${{ matrix.features }}
+
+  test-all-features:
+    name: Tests almost all features
+    runs-on: ubuntu-22.04
+    if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
+    steps:
+      - uses: actions/checkout@v5
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
         run: |
-          apt-get update
+          sudo rm -rf "/opt/ghc" || true
-          apt-get install --assume-yes build-essential curl
+          sudo rm -rf "/usr/share/dotnet" || true
-      - uses: dtolnay/rust-toolchain@1.89
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Run cargo build with almost all features
         run: |
-          cargo build --workspace --locked --release --features "$(cargo xtask list-features --exclude-feature cuda,test-ollama)"
+          cargo build --workspace --locked --features "$(cargo xtask list-features --exclude-feature cuda,test-ollama)"
       - name: Run cargo test with almost all features
         run: |
-          cargo test --workspace --locked --release --features "$(cargo xtask list-features --exclude-feature cuda,test-ollama)"
+          cargo test --workspace --locked --features "$(cargo xtask list-features --exclude-feature cuda,test-ollama)"
 
   ollama-ubuntu:
     name: Test with Ollama
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
-    strategy:
-      matrix:
-        features: ["", "--features enterprise"]
     env:
       MEILI_TEST_OLLAMA_SERVER: "http://localhost:11434"
     steps:
       - uses: actions/checkout@v5
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
       - name: Install Ollama
         run: |
           curl -fsSL https://ollama.com/install.sh | sudo -E sh
@@ -123,21 +156,21 @@ jobs:
         uses: actions-rs/cargo@v1
         with:
           command: test
-          args: --locked --release --all --features test-ollama ollama ${{ matrix.features }}
+          args: --locked -p meilisearch --features test-ollama ollama
 
   test-disabled-tokenization:
     name: Test disabled tokenization
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
-    container:
-      image: ubuntu:22.04
     if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
     steps:
       - uses: actions/checkout@v5
-      - name: Install needed dependencies
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
         run: |
-          apt-get update
+          sudo rm -rf "/opt/ghc" || true
-          apt-get install --assume-yes build-essential curl
+          sudo rm -rf "/usr/share/dotnet" || true
-      - uses: dtolnay/rust-toolchain@1.89
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Run cargo tree without default features and check lindera is not present
         run: |
           if cargo tree -f '{p} {f}' -e normal --no-default-features | grep -qz lindera; then
@@ -148,35 +181,39 @@ jobs:
         run: |
           cargo tree -f '{p} {f}' -e normal | grep lindera -qz
 
-  # We run tests in debug also, to make sure that the debug_assertions are hit
+  build:
-  test-debug:
+    name: Build in release
-    name: Run tests in debug
+    runs-on: ubuntu-22.04
+    steps:
+      - uses: actions/checkout@v5
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
+      - name: Cache dependencies
+        uses: Swatinem/rust-cache@v2.8.0
+      - name: Build
+        run: cargo build --release --locked --target x86_64-unknown-linux-gnu
+
+  clippy:
+    name: Run Clippy
     runs-on: ubuntu-22.04
     strategy:
       matrix:
         features: ["", "--features enterprise"]
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
-      - name: Cache dependencies
+        run: |
-        uses: Swatinem/rust-cache@v2.8.0
+          sudo rm -rf "/opt/ghc" || true
-      - name: Run tests in debug
+          sudo rm -rf "/usr/share/dotnet" || true
-        uses: actions-rs/cargo@v1
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
-          command: test
-          args: --locked --all ${{ matrix.features }}
-
-  clippy:
-    name: Run Clippy
-    runs-on: ubuntu-latest
-    strategy:
-      matrix:
-        features: ["", "--features enterprise"]
-    steps:
-      - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
-        with:
-          profile: minimal
           components: clippy
       - name: Cache dependencies
         uses: Swatinem/rust-cache@v2.8.0
@@ -184,18 +221,21 @@ jobs:
         uses: actions-rs/cargo@v1
         with:
           command: clippy
-          args: --all-targets ${{ matrix.features }} -- --deny warnings
+          args: --all-targets ${{ matrix.features }} -- --deny warnings -D clippy::todo
 
   fmt:
     name: Run Rustfmt
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
         with:
-          profile: minimal
-          toolchain: nightly-2024-07-09
-          override: true
           components: rustfmt
       - name: Cache dependencies
         uses: Swatinem/rust-cache@v2.8.0
@@ -206,3 +246,23 @@ jobs:
         run: |
           echo -ne "\n" > crates/benchmarks/benches/datasets_paths.rs
           cargo fmt --all -- --check
+
+  declarative-tests:
+    name: Run declarative tests
+    runs-on: ubuntu-22.04-arm
+    permissions:
+      contents: read
+    steps:
+      - uses: actions/checkout@v5
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
+        run: |
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
+      - name: Cache dependencies
+        uses: Swatinem/rust-cache@v2.8.0
+      - name: Run declarative tests
+        run: |
+          cargo xtask test workloads/tests/*.json

.github/workflows/update-cargo-toml-version.yml

@@ -18,9 +18,13 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v5
-      - uses: dtolnay/rust-toolchain@1.89
+      - name: Clean space as per https://github.com/actions/virtual-environments/issues/709
-        with:
+        run: |
-          profile: minimal
+          sudo rm -rf "/opt/ghc" || true
+          sudo rm -rf "/usr/share/dotnet" || true
+          sudo rm -rf "/usr/local/lib/android" || true
+          sudo rm -rf "/usr/local/share/boost" || true
+      - uses: dtolnay/rust-toolchain@1.91.1
       - name: Install sd
         run: cargo install sd
       - name: Update Cargo.toml file

BENCHMARKS.md

@@ -124,6 +124,7 @@ They are JSON files with the following structure (comments are not actually supp
 {
   // Name of the workload. Must be unique to the workload, as it will be used to group results on the dashboard.
   "name": "hackernews.ndjson_1M,no-threads",
+  "type": "bench",
   // Number of consecutive runs of the commands that should be performed.
   // Each run uses a fresh instance of Meilisearch and a fresh database.
   // Each run produces its own report file.

Cargo.lock

@@ -580,7 +580,7 @@ source = "git+https://github.com/meilisearch/bbqueue#e8af4a4bccc8eb36b2b0442c4a9
 
 [[package]]
 name = "benchmarks"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "bumpalo",
@@ -790,11 +790,11 @@ dependencies = [
 
 [[package]]
 name = "build-info"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "time",
- "vergen-git2",
+ "vergen-gitcl",
 ]
 
 [[package]]
@@ -1786,7 +1786,7 @@ dependencies = [
 
 [[package]]
 name = "dump"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "big_s",
@@ -2018,7 +2018,7 @@ checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
 
 [[package]]
 name = "file-store"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "tempfile",
  "thiserror 2.0.17",
@@ -2040,7 +2040,7 @@ dependencies = [
 
 [[package]]
 name = "filter-parser"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "insta",
  "levenshtein_automata",
@@ -2068,7 +2068,7 @@ dependencies = [
 
 [[package]]
 name = "flatten-serde-json"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "criterion",
  "serde_json",
@@ -2231,7 +2231,7 @@ dependencies = [
 
 [[package]]
 name = "fuzzers"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "arbitrary",
  "bumpalo",
@@ -2604,19 +2604,6 @@ version = "0.32.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e629b9b98ef3dd8afe6ca2bd0f89306cec16d43d907889945bc5d6687f2f13c7"
 
-[[package]]
-name = "git2"
-version = "0.20.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2deb07a133b1520dc1a5690e9bd08950108873d7ed5de38dcc74d3b5ebffa110"
-dependencies = [
- "bitflags 2.10.0",
- "libc",
- "libgit2-sys",
- "log",
- "url",
-]
-
 [[package]]
 name = "glob"
 version = "0.3.3"
@@ -2711,9 +2698,9 @@ dependencies = [
 
 [[package]]
 name = "hannoy"
-version = "0.0.9-nested-rtxns-2"
+version = "0.1.0-nested-rtxns"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "06eda090938d9dcd568c8c2a5de383047ed9191578ebf4a342d2975d16e621f2"
+checksum = "be82bf3f2108ddc8885e3d306fcd7f4692066bfe26065ca8b42ba417f3c26dd1"
 dependencies = [
  "bytemuck",
  "byteorder",
@@ -3198,7 +3185,7 @@ dependencies = [
 
 [[package]]
 name = "index-scheduler"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "backoff",
@@ -3216,6 +3203,7 @@ dependencies = [
  "enum-iterator",
  "file-store",
  "flate2",
+ "hashbrown 0.15.5",
  "indexmap",
  "insta",
  "maplit",
@@ -3238,6 +3226,7 @@ dependencies = [
  "tokio",
  "tracing",
  "ureq",
+ "urlencoding",
  "uuid",
 ]
 
@@ -3460,7 +3449,7 @@ dependencies = [
 
 [[package]]
 name = "json-depth-checker"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "criterion",
  "serde_json",
@@ -3557,18 +3546,6 @@ dependencies = [
  "rle-decode-fast",
 ]
 
-[[package]]
-name = "libgit2-sys"
-version = "0.18.2+1.9.1"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1c42fe03df2bd3c53a3a9c7317ad91d80c81cd1fb0caec8d7cc4cd2bfa10c222"
-dependencies = [
- "cc",
- "libc",
- "libz-sys",
- "pkg-config",
-]
-
 [[package]]
 name = "libloading"
 version = "0.8.9"
@@ -3626,18 +3603,6 @@ dependencies = [
  "zlib-rs",
 ]
 
-[[package]]
-name = "libz-sys"
-version = "1.1.22"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8b70e7a7df205e92a1a4cd9aaae7898dac0aa555503cc0a649494d0d60e7651d"
-dependencies = [
- "cc",
- "libc",
- "pkg-config",
- "vcpkg",
-]
-
 [[package]]
 name = "lindera"
 version = "0.43.3"
@@ -3974,7 +3939,7 @@ checksum = "ae960838283323069879657ca3de837e9f7bbb4c7bf6ea7f1b290d5e9476d2e0"
 
 [[package]]
 name = "meili-snap"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "insta",
  "md5 0.8.0",
@@ -3985,7 +3950,7 @@ dependencies = [
 
 [[package]]
 name = "meilisearch"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "actix-cors",
  "actix-http",
@@ -4083,7 +4048,7 @@ dependencies = [
 
 [[package]]
 name = "meilisearch-auth"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "base64 0.22.1",
  "enum-iterator",
@@ -4102,10 +4067,11 @@ dependencies = [
 
 [[package]]
 name = "meilisearch-types"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "actix-web",
  "anyhow",
+ "base64 0.22.1",
  "bumpalo",
  "bumparaw-collections",
  "byte-unit",
@@ -4118,6 +4084,7 @@ dependencies = [
  "flate2",
  "fst",
  "insta",
+ "itertools 0.14.0",
  "meili-snap",
  "memmap2",
  "milli",
@@ -4131,13 +4098,14 @@ dependencies = [
  "thiserror 2.0.17",
  "time",
  "tokio",
+ "urlencoding",
  "utoipa",
  "uuid",
 ]
 
 [[package]]
 name = "meilitool"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "clap",
@@ -4171,7 +4139,7 @@ dependencies = [
 
 [[package]]
 name = "milli"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "arroy",
  "bbqueue",
@@ -4750,7 +4718,7 @@ checksum = "9b4f627cb1b25917193a259e49bdad08f671f8d9708acfd5fe0a8c1455d87220"
 
 [[package]]
 name = "permissive-json-pointer"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "big_s",
  "serde_json",
@@ -6072,6 +6040,20 @@ name = "similar"
 version = "2.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bbbb5d9659141646ae647b42fe094daf6c6192d1620870b449d9557f748b2daa"
+dependencies = [
+ "bstr",
+ "unicode-segmentation",
+]
+
+[[package]]
+name = "similar-asserts"
+version = "1.7.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b5b441962c817e33508847a22bd82f03a30cff43642dc2fae8b050566121eb9a"
+dependencies = [
+ "console",
+ "similar",
+]
 
 [[package]]
 name = "simple_asn1"
@@ -7105,12 +7087,6 @@ version = "0.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ba73ea9cf16a25df0c8caa16c51acb937d5712a8429db78a3ee29d5dcacd3a65"
 
-[[package]]
-name = "vcpkg"
-version = "0.2.15"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "accd4ea62f7bb7a82fe23066fb0957d48ef677f6eeb8215f372f52e48bb32426"
-
 [[package]]
 name = "vergen"
 version = "9.0.6"
@@ -7124,14 +7100,13 @@ dependencies = [
 ]
 
 [[package]]
-name = "vergen-git2"
+name = "vergen-gitcl"
-version = "1.0.7"
+version = "1.0.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4f6ee511ec45098eabade8a0750e76eec671e7fb2d9360c563911336bea9cac1"
+checksum = "b9dfc1de6eb2e08a4ddf152f1b179529638bedc0ea95e6d667c014506377aefe"
 dependencies = [
  "anyhow",
  "derive_builder",
- "git2",
  "rustversion",
  "time",
  "vergen",
@@ -7783,7 +7758,7 @@ dependencies = [
 
 [[package]]
 name = "xtask"
-version = "1.28.2"
+version = "1.30.0"
 dependencies = [
  "anyhow",
  "build-info",
@@ -7792,9 +7767,11 @@ dependencies = [
  "futures-core",
  "futures-util",
  "reqwest",
+ "semver",
  "serde",
  "serde_json",
  "sha2",
+ "similar-asserts",
  "sysinfo",
  "time",
  "tokio",

Cargo.toml

@@ -23,7 +23,7 @@ members = [
 ]
 
 [workspace.package]
-version = "1.28.2"
+version = "1.30.0"
 authors = [
     "Quentin de Quelen <quentin@dequelen.me>",
     "Clément Renault <clement@meilisearch.com>",

README.md

@@ -104,6 +104,7 @@ Meilisearch is available in two editions:
 
 - Includes advanced features such as:
   - Sharding
+  - S3-streaming snapshots
 - Governed by a [commercial license](./LICENSE-EE) or the [Business Source License 1.1](https://mariadb.com/bsl11)
 - Not allowed in production without a commercial agreement with Meilisearch.
   - You may use, modify, and distribute the Licensed Work for non-production purposes only, such as testing, development, or evaluation.

TESTING.md

@@ -0,0 +1,326 @@
+# Declarative tests
+
+Declarative tests ensure that Meilisearch features remain stable across versions.
+
+While we already have unit tests, those are run against **temporary databases** that are created fresh each time and therefore never risk corruption.
+
+Declarative tests instead **simulate the lifetime of a database**: they chain together commands and requests to change the binary, verifying that database state and API responses remain consistent.
+
+## Basic example
+
+```jsonc
+{
+  "type": "test",
+  "name": "api-keys",
+  "binary": { // the first command will run on the binary following this specification.
+    "source": "release", // get the binary as a release from GitHub
+    "version": "1.19.0", // version to fetch
+    "edition": "community" // edition to fetch
+  },
+  "commands": []
+}
+```
+
+This example defines a no-op test (it does nothing).
+
+If the file is saved at `workloads/tests/example.json`, you can run it with:
+
+```bash
+cargo xtask test workloads/tests/example.json
+```
+
+## Commands
+
+Commands represent API requests sent to Meilisearch endpoints during a test.
+
+They are executed sequentially, and their responses can be validated to ensure consistent behavior across upgrades.
+
+```jsonc
+
+{
+  "route": "keys",
+  "method": "POST",
+  "body": {
+    "inline": {
+      "actions": [
+        "search",
+        "documents.add"
+      ],
+      "description": "Test API Key",
+      "expiresAt": null,
+      "indexes": [ "movies" ]
+    }
+  }
+}
+```
+
+This command issues a `POST /keys` request, creating an API key with permissions to search and add documents in the `movies` index.
+
+### Using assets in commands
+
+To keep tests concise and reusable, you can define **assets** at the root of the workload file.
+
+Assets are external data sources (such as datasets) that are cached between runs, making tests faster and easier to read.
+
+```jsonc
+{
+  "type": "test",
+  "name": "movies",
+  "binary": {
+    "source": "release",
+    "version": "1.19.0",
+    "edition": "community"
+  },
+  "assets": {
+    "movies.json": {
+      "local_location": null,
+      "remote_location": "https://milli-benchmarks.fra1.digitaloceanspaces.com/bench/datasets/movies.json",
+      "sha256": "5b6e4cb660bc20327776e8a33ea197b43d9ec84856710ead1cc87ab24df77de1"
+    }
+  },
+  "commands": [
+    {
+      "route": "indexes/movies/documents",
+      "method": "POST",
+      "body": {
+        "asset": "movies.json"
+      }
+    }
+  ]
+}
+```
+
+In this example:
+- The `movies.json` dataset is defined as an asset, pointing to a remote URL.
+- The SHA-256 checksum ensures integrity.
+- The `POST /indexes/movies/documents` command uses this asset as the request body.
+
+This makes the test much cleaner than inlining a large dataset directly into the command.
+
+For asset handling, please refer to the [declarative benchmarks documentation](/BENCHMARKS.md#adding-new-assets).
+
+### Asserting responses
+
+Commands can specify both the **expected status code** and the **expected response body**.
+
+```jsonc
+{
+  "route": "indexes/movies/documents",
+  "method": "POST",
+  "body": {
+    "asset": "movies.json"
+  },
+  "expectedStatus": 202,
+  "expectedResponse": {
+    "enqueuedAt": "[timestamp]", // Set to a bracketed string to ignore the value
+    "indexUid": "movies",
+    "status": "enqueued",
+    "taskUid": 1,
+    "type": "documentAdditionOrUpdate"
+  },
+  "synchronous": "WaitForTask"
+}
+```
+
+Manually writing `expectedResponse` fields can be tedious.
+
+Instead, you can let the test runner populate them automatically:
+
+```bash
+# Run the workload to populate expected fields. Only adds the missing ones, doesn't change existing data
+cargo xtask test workloads/tests/example.json --add-missing-responses
+
+# OR
+
+# Run the workload to populate expected fields. Updates all fields including existing ones
+cargo xtask test workloads/tests/example.json --update-responses
+```
+
+This workflow is recommended:
+
+1. Write the test without expected fields.
+2. Run it with `--add-missing-responses` to capture the actual responses.
+3. Review and commit the generated expectations.
+
+## Changing binary
+
+It is possible to insert an instruction to change the current Meilisearch instance from one binary specification to another during a test.
+
+When executed, such an instruction will:
+1. Stop the current Meilisearch instance.
+2. Fetch the binary specified by the instruction.
+3. Restart the server with the specified binary on the same database.
+
+```jsonc
+{
+  "type": "test",
+  "name": "movies",
+  "binary": {
+    "source": "release",
+    "version": "1.19.0", // start with version v1.19.0
+    "edition": "community"
+  },
+  "assets": {
+    "movies.json": {
+      "local_location": null,
+      "remote_location": "https://milli-benchmarks.fra1.digitaloceanspaces.com/bench/datasets/movies.json",
+      "sha256": "5b6e4cb660bc20327776e8a33ea197b43d9ec84856710ead1cc87ab24df77de1"
+    }
+  },
+  "commands": [
+    // setup some data
+    {
+      "route": "indexes/movies/documents",
+      "method": "POST",
+      "body": {
+        "asset": "movies.json"
+      }
+    },
+    // switch binary to v1.24.0
+    {
+      "binary": {
+        "source": "release",
+        "version": "1.24.0",
+        "edition": "community"
+      }
+    }
+  ]
+}
+```
+
+### Typical Usage
+
+In most cases, the change binary instruction will be used to update a database.
+
+- **Set up** some data using commands on an older version.
+- **Upgrade** to the latest version.
+- **Assert** that the data and API behavior remain correct after the upgrade.
+
+To properly test the dumpless upgrade, one should typically:
+
+1. Open the database without processing the update task: Use a `binary` instruction to switch to the desired version, passing `--experimental-dumpless-upgrade` and `--experimental-max-number-of-batched-tasks=0` as extra CLI arguments
+2. Check that the search, stats and task queue still work.
+3. Open the database and process the update task: Use a `binary` instruction to switch to the desired version, passing `--experimental-dumpless-upgrade` as the extra CLI argument. Use a `health` command to wait for the upgrade task to finish.
+4. Check that the indexing, search, stats, and task queue still work.
+
+```jsonc
+{
+  "type": "test",
+  "name": "movies",
+  "binary": {
+    "source": "release",
+    "version": "1.12.0",
+    "edition": "community"
+  },
+  "commands": [
+    // 0. Run commands to populate the database
+    {
+      // ..
+    },
+    // 1. Open the database with new MS without processing the update task
+    {
+      "binary": {
+        "source": "build", // build the binary from the sources in the current git repository
+        "edition": "community",
+        "extraCliArgs": [
+          "--experimental-dumpless-upgrade", // allows to open with a newer MS
+          "--experimental-max-number-of-batched-tasks=0" // prevent processing of the update task
+        ]
+      }
+    },
+    // 2. Check the search etc.
+    {
+      // ..
+    },
+    // 3. Open the database with new MS and processing the update task
+    {
+      "binary": {
+        "source": "build", // build the binary from the sources in the current git repository
+        "edition": "community",
+        "extraCliArgs": [
+          "--experimental-dumpless-upgrade" // allows to open with a newer MS
+          // no `--experimental-max-number-of-batched-tasks=0`
+        ]
+      }
+    },
+    // 4. Check the indexing, search, etc.
+    {
+      // ..
+    }
+  ]
+}
+```
+
+This ensures backward compatibility: databases created with older Meilisearch versions should remain functional and consistent after an upgrade.
+
+## Variables
+
+Sometimes a command needs to use a value returned by a **previous response**.
+These values can be captured and reused using the register field.
+
+```jsonc
+{
+  "route": "keys",
+  "method": "POST",
+  "body": {
+    "inline": {
+        "actions": [
+        "search",
+        "documents.add"
+        ],
+        "description": "Test API Key",
+        "expiresAt": null,
+        "indexes": [ "movies" ]
+    }
+  },
+  "expectedResponse": {
+      "key": "c6f64630bad2996b1f675007c8800168e14adf5d6a7bb1a400a6d2b158050eaf",
+      // ...
+  },
+  "register": {
+    "key": "/key"
+  },
+  "synchronous": "WaitForResponse"
+}
+```
+
+The `register` field captures the value at the JSON path `/key` from the response.
+Paths follow the **JavaScript Object Notation Pointer (RFC 6901)** format.
+Registered variables are available for all subsequent commands.
+
+Registered variables can be referenced by wrapping their name in double curly braces:
+
+In the route/path:
+
+```jsonc
+{
+  "route": "tasks/{{ task_id }}",
+  "method": "GET"
+}
+```
+
+In the request body:
+
+```jsonc
+{
+  "route": "indexes/movies/documents",
+  "method": "PATCH",
+  "body": {
+    "inline": {
+      "id": "{{ document_id }}",
+      "overview": "Shazam turns evil and the world is in danger.",
+    }
+  }
+}
+```
+
+Or they can be referenced by their name (**without curly braces**) as an API key:
+
+```jsonc
+{
+  "route": "indexes/movies/documents",
+  "method": "POST",
+  "body": { /* ... */ },
+  "apiKeyVariable": "key" // The **content** of the key variable will be used as an API key
+}
+```

crates/benchmarks/benches/indexing.rs

@@ -10,7 +10,7 @@ use milli::documents::PrimaryKey;
 use milli::heed::{EnvOpenOptions, RwTxn};
 use milli::progress::Progress;
 use milli::update::new::indexer;
-use milli::update::{IndexerConfig, Settings};
+use milli::update::{IndexerConfig, MissingDocumentPolicy, Settings};
 use milli::vector::RuntimeEmbedders;
 use milli::{FilterableAttributesRule, Index};
 use rand::seq::SliceRandom;
@@ -21,6 +21,10 @@ use roaring::RoaringBitmap;
 #[global_allocator]
 static ALLOC: mimalloc::MiMalloc = mimalloc::MiMalloc;
 
+fn no_cancel() -> bool {
+    false
+}
+
 const BENCHMARK_ITERATION: usize = 10;
 
 fn setup_dir(path: impl AsRef<Path>) {
@@ -65,7 +69,7 @@ fn setup_settings<'t>(
     let sortable_fields = sortable_fields.iter().map(|s| s.to_string()).collect();
     builder.set_sortable_fields(sortable_fields);
 
-    builder.execute(&|| false, &Progress::default(), Default::default()).unwrap();
+    builder.execute(&no_cancel, &Progress::default(), Default::default()).unwrap();
 }
 
 fn setup_index_with_settings(
@@ -142,7 +146,7 @@ fn indexing_songs_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -152,7 +156,7 @@ fn indexing_songs_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -168,7 +172,7 @@ fn indexing_songs_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -210,7 +214,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -220,7 +224,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -236,7 +240,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -256,7 +260,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -266,7 +270,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -282,7 +286,7 @@ fn reindexing_songs_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -326,7 +330,7 @@ fn deleting_songs_in_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -336,7 +340,7 @@ fn deleting_songs_in_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -352,7 +356,7 @@ fn deleting_songs_in_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -404,7 +408,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS_1_2, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -414,7 +418,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -430,7 +434,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -450,7 +454,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS_3_4, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -460,7 +464,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -476,7 +480,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -492,7 +496,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS_4_4, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -502,7 +506,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -518,7 +522,7 @@ fn indexing_songs_in_three_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -561,7 +565,7 @@ fn indexing_songs_without_faceted_numbers(c: &mut Criterion) {
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
 
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -571,7 +575,7 @@ fn indexing_songs_without_faceted_numbers(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -587,7 +591,7 @@ fn indexing_songs_without_faceted_numbers(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -629,7 +633,7 @@ fn indexing_songs_without_faceted_fields(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_SONGS, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -639,7 +643,7 @@ fn indexing_songs_without_faceted_fields(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -655,7 +659,7 @@ fn indexing_songs_without_faceted_fields(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -697,7 +701,7 @@ fn indexing_wiki(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -707,7 +711,7 @@ fn indexing_wiki(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -723,7 +727,7 @@ fn indexing_wiki(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -764,7 +768,7 @@ fn reindexing_wiki(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -774,7 +778,7 @@ fn reindexing_wiki(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -790,7 +794,7 @@ fn reindexing_wiki(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -810,7 +814,7 @@ fn reindexing_wiki(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -820,7 +824,7 @@ fn reindexing_wiki(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -836,7 +840,7 @@ fn reindexing_wiki(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -879,7 +883,7 @@ fn deleting_wiki_in_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -889,7 +893,7 @@ fn deleting_wiki_in_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -905,7 +909,7 @@ fn deleting_wiki_in_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -957,7 +961,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents =
                     utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_1_2, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -967,7 +971,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -983,7 +987,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1004,7 +1008,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents =
                     utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_3_4, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1014,7 +1018,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1030,7 +1034,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1047,7 +1051,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents =
                     utils::documents_from(datasets_paths::SMOL_WIKI_ARTICLES_4_4, "csv");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1057,7 +1061,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1073,7 +1077,7 @@ fn indexing_wiki_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1115,7 +1119,7 @@ fn indexing_movies_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1125,7 +1129,7 @@ fn indexing_movies_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1141,7 +1145,7 @@ fn indexing_movies_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1182,7 +1186,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1192,7 +1196,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1208,7 +1212,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1228,7 +1232,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1238,7 +1242,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1254,7 +1258,7 @@ fn reindexing_movies_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1297,7 +1301,7 @@ fn deleting_movies_in_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1307,7 +1311,7 @@ fn deleting_movies_in_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1323,7 +1327,7 @@ fn deleting_movies_in_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1372,7 +1376,7 @@ fn delete_documents_from_ids(index: Index, document_ids_to_delete: Vec<RoaringBi
             Some(primary_key),
             &document_changes,
             RuntimeEmbedders::default(),
-            &|| false,
+            &no_cancel,
             &Progress::default(),
             &Default::default(),
         )
@@ -1412,7 +1416,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES_1_2, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1422,7 +1426,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1438,7 +1442,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1458,7 +1462,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES_3_4, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1468,7 +1472,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1484,7 +1488,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1500,7 +1504,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::MOVIES_4_4, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1510,7 +1514,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1526,7 +1530,7 @@ fn indexing_movies_in_three_batches(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1591,7 +1595,7 @@ fn indexing_nested_movies_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::NESTED_MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1601,7 +1605,7 @@ fn indexing_nested_movies_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1617,7 +1621,7 @@ fn indexing_nested_movies_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1683,7 +1687,7 @@ fn deleting_nested_movies_in_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::NESTED_MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1693,7 +1697,7 @@ fn deleting_nested_movies_in_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1709,7 +1713,7 @@ fn deleting_nested_movies_in_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1767,7 +1771,7 @@ fn indexing_nested_movies_without_faceted_fields(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::NESTED_MOVIES, "json");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1777,7 +1781,7 @@ fn indexing_nested_movies_without_faceted_fields(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1793,7 +1797,7 @@ fn indexing_nested_movies_without_faceted_fields(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1835,7 +1839,7 @@ fn indexing_geo(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_ALL_COUNTRIES, "jsonl");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1845,7 +1849,7 @@ fn indexing_geo(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1861,7 +1865,7 @@ fn indexing_geo(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1902,7 +1906,7 @@ fn reindexing_geo(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_ALL_COUNTRIES, "jsonl");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1912,7 +1916,7 @@ fn reindexing_geo(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1928,7 +1932,7 @@ fn reindexing_geo(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -1948,7 +1952,7 @@ fn reindexing_geo(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_ALL_COUNTRIES, "jsonl");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -1958,7 +1962,7 @@ fn reindexing_geo(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -1974,7 +1978,7 @@ fn reindexing_geo(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )
@@ -2017,7 +2021,7 @@ fn deleting_geo_in_batches_default(c: &mut Criterion) {
 
                 let mut indexer = indexer::DocumentOperation::new();
                 let documents = utils::documents_from(datasets_paths::SMOL_ALL_COUNTRIES, "jsonl");
-                indexer.replace_documents(&documents).unwrap();
+                indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
                 let indexer_alloc = Bump::new();
                 let (document_changes, _operation_stats, primary_key) = indexer
@@ -2027,7 +2031,7 @@ fn deleting_geo_in_batches_default(c: &mut Criterion) {
                         &rtxn,
                         None,
                         &mut new_fields_ids_map,
-                        &|| false,
+                        &no_cancel,
                         Progress::default(),
                         None,
                     )
@@ -2043,7 +2047,7 @@ fn deleting_geo_in_batches_default(c: &mut Criterion) {
                     primary_key,
                     &document_changes,
                     RuntimeEmbedders::default(),
-                    &|| false,
+                    &no_cancel,
                     &Progress::default(),
                     &Default::default(),
                 )

crates/benchmarks/benches/utils.rs

@@ -13,7 +13,7 @@ use milli::documents::sort::recursive_sort;
 use milli::heed::EnvOpenOptions;
 use milli::progress::Progress;
 use milli::update::new::indexer;
-use milli::update::{IndexerConfig, Settings};
+use milli::update::{IndexerConfig, MissingDocumentPolicy, Settings};
 use milli::vector::RuntimeEmbedders;
 use milli::{Criterion, Filter, Index, Object, TermsMatchingStrategy};
 use serde_json::Value;
@@ -111,7 +111,7 @@ pub fn base_setup(conf: &Conf) -> Index {
 
     let documents = documents_from(conf.dataset, conf.dataset_format);
     let mut indexer = indexer::DocumentOperation::new();
-    indexer.replace_documents(&documents).unwrap();
+    indexer.replace_documents(&documents, MissingDocumentPolicy::default()).unwrap();
 
     let indexer_alloc = Bump::new();
     let (document_changes, _operation_stats, primary_key) = indexer

crates/build-info/Cargo.toml

@@ -15,4 +15,4 @@ time = { version = "0.3.44", features = ["parsing"] }
 
 [build-dependencies]
 anyhow = "1.0.100"
-vergen-git2 = "1.0.7"
+vergen-gitcl = "1.0.8"

crates/build-info/build.rs

@@ -15,7 +15,7 @@ fn emit_git_variables() -> anyhow::Result<()> {
     // Note: any code that needs VERGEN_ environment variables should take care to define them manually in the Dockerfile and pass them
     // in the corresponding GitHub workflow (publish_docker.yml).
     // This is due to the Dockerfile building the binary outside of the git directory.
-    let mut builder = vergen_git2::Git2Builder::default();
+    let mut builder = vergen_gitcl::GitclBuilder::default();
 
     builder.branch(true);
     builder.commit_timestamp(true);
@@ -25,5 +25,5 @@ fn emit_git_variables() -> anyhow::Result<()> {
 
     let git2 = builder.build()?;
 
-    vergen_git2::Emitter::default().fail_on_error().add_instructions(&git2)?.emit()
+    vergen_gitcl::Emitter::default().fail_on_error().add_instructions(&git2)?.emit()
 }

crates/build-info/src/main.rs

@@ -0,0 +1,6 @@
+use build_info::BuildInfo;
+
+fn main() {
+    let info = BuildInfo::from_build();
+    dbg!(info);
+}

crates/dump/src/lib.rs

@@ -9,8 +9,9 @@ use meilisearch_types::error::ResponseError;
 use meilisearch_types::keys::Key;
 use meilisearch_types::milli::update::IndexDocumentsMethod;
 use meilisearch_types::settings::Unchecked;
+use meilisearch_types::tasks::network::{DbTaskNetwork, NetworkTopologyChange};
 use meilisearch_types::tasks::{
-    Details, ExportIndexSettings, IndexSwap, KindWithContent, Status, Task, TaskId, TaskNetwork,
+    Details, ExportIndexSettings, IndexSwap, KindWithContent, Status, Task, TaskId,
 };
 use meilisearch_types::InstanceUid;
 use roaring::RoaringBitmap;
@@ -95,7 +96,7 @@ pub struct TaskDump {
     )]
     pub finished_at: Option<OffsetDateTime>,
     #[serde(default, skip_serializing_if = "Option::is_none")]
-    pub network: Option<TaskNetwork>,
+    pub network: Option<DbTaskNetwork>,
     #[serde(default, skip_serializing_if = "Option::is_none")]
     pub custom_metadata: Option<String>,
 }
@@ -163,6 +164,7 @@ pub enum KindDump {
     IndexCompaction {
         index_uid: String,
     },
+    NetworkTopologyChange(NetworkTopologyChange),
 }
 
 impl From<Task> for TaskDump {
@@ -249,6 +251,9 @@ impl From<KindWithContent> for KindDump {
             KindWithContent::IndexCompaction { index_uid } => {
                 KindDump::IndexCompaction { index_uid }
             }
+            KindWithContent::NetworkTopologyChange(network_topology_change) => {
+                KindDump::NetworkTopologyChange(network_topology_change)
+            }
         }
     }
 }
@@ -560,7 +565,8 @@ pub(crate) mod test {
         Network {
             local: Some("myself".to_string()),
             remotes: maplit::btreemap! {"other".to_string() => Remote { url: "http://test".to_string(), search_api_key: Some("apiKey".to_string()), write_api_key: Some("docApiKey".to_string()) }},
-            sharding: false,
+            leader: None,
+            version: Default::default(),
         }
     }
 
@@ -614,7 +620,10 @@ pub(crate) mod test {
         assert_eq!(dump.features().unwrap().unwrap(), expected);
 
         // ==== checking the network
-        let expected = create_test_network();
+        let mut expected = create_test_network();
+        // from v1.29, we drop `leader` and `local` on import
+        expected.leader = None;
+        expected.local = None;
         assert_eq!(&expected, dump.network().unwrap().unwrap());
     }
 }

crates/dump/src/reader/mod.rs

@@ -434,7 +434,11 @@ pub(crate) mod test {
         // network
 
         let network = dump.network().unwrap().unwrap();
-        insta::assert_snapshot!(network.local.as_ref().unwrap(), @"ms-0");
+
+        // since v1.29 we are dropping `local` and `leader` on import
+        insta::assert_snapshot!(network.local.is_none(), @"true");
+        insta::assert_snapshot!(network.leader.is_none(), @"true");
+
         insta::assert_snapshot!(network.remotes.get("ms-0").as_ref().unwrap().url, @"http://localhost:7700");
         insta::assert_snapshot!(network.remotes.get("ms-0").as_ref().unwrap().search_api_key.is_none(), @"true");
         insta::assert_snapshot!(network.remotes.get("ms-1").as_ref().unwrap().url, @"http://localhost:7701");

crates/dump/src/reader/v2/settings.rs

@@ -107,19 +107,14 @@ impl Settings<Unchecked> {
     }
 }
 
-#[derive(Debug, Clone, PartialEq)]
+#[derive(Default, Debug, Clone, PartialEq)]
 pub enum Setting<T> {
     Set(T),
     Reset,
+    #[default]
     NotSet,
 }
 
-impl<T> Default for Setting<T> {
-    fn default() -> Self {
-        Self::NotSet
-    }
-}
-
 impl<T> Setting<T> {
     pub const fn is_not_set(&self) -> bool {
         matches!(self, Self::NotSet)

crates/dump/src/reader/v3/settings.rs

@@ -161,19 +161,14 @@ pub struct Facets {
     pub min_level_size: Option<NonZeroUsize>,
 }
 
-#[derive(Debug, Clone, PartialEq, Eq)]
+#[derive(Default, Debug, Clone, PartialEq, Eq)]
 pub enum Setting<T> {
     Set(T),
     Reset,
+    #[default]
     NotSet,
 }
 
-impl<T> Default for Setting<T> {
-    fn default() -> Self {
-        Self::NotSet
-    }
-}
-
 impl<T> Setting<T> {
     pub fn map<U, F>(self, f: F) -> Setting<U>
     where

crates/dump/src/reader/v4/meta.rs

@@ -1,9 +1,7 @@
 use std::fmt::{self, Display, Formatter};
-use std::marker::PhantomData;
 use std::str::FromStr;
 
-use serde::de::Visitor;
+use serde::Deserialize;
-use serde::{Deserialize, Deserializer};
 use uuid::Uuid;
 
 use super::settings::{Settings, Unchecked};
@@ -82,59 +80,3 @@ impl Display for IndexUidFormatError {
 }
 
 impl std::error::Error for IndexUidFormatError {}
-
-/// A type that tries to match either a star (*) or
-/// any other thing that implements `FromStr`.
-#[derive(Debug)]
-#[cfg_attr(test, derive(serde::Serialize))]
-pub enum StarOr<T> {
-    Star,
-    Other(T),
-}
-
-impl<'de, T, E> Deserialize<'de> for StarOr<T>
-where
-    T: FromStr<Err = E>,
-    E: Display,
-{
-    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
-    where
-        D: Deserializer<'de>,
-    {
-        /// Serde can't differentiate between `StarOr::Star` and `StarOr::Other` without a tag.
-        /// Simply using `#[serde(untagged)]` + `#[serde(rename="*")]` will lead to attempting to
-        /// deserialize everything as a `StarOr::Other`, including "*".
-        /// [`#[serde(other)]`](https://serde.rs/variant-attrs.html#other) might have helped but is
-        /// not supported on untagged enums.
-        struct StarOrVisitor<T>(PhantomData<T>);
-
-        impl<T, FE> Visitor<'_> for StarOrVisitor<T>
-        where
-            T: FromStr<Err = FE>,
-            FE: Display,
-        {
-            type Value = StarOr<T>;
-
-            fn expecting(&self, formatter: &mut Formatter) -> std::fmt::Result {
-                formatter.write_str("a string")
-            }
-
-            fn visit_str<SE>(self, v: &str) -> Result<Self::Value, SE>
-            where
-                SE: serde::de::Error,
-            {
-                match v {
-                    "*" => Ok(StarOr::Star),
-                    v => {
-                        let other = FromStr::from_str(v).map_err(|e: T::Err| {
-                            SE::custom(format!("Invalid `other` value: {}", e))
-                        })?;
-                        Ok(StarOr::Other(other))
-                    }
-                }
-            }
-        }
-
-        deserializer.deserialize_str(StarOrVisitor(PhantomData))
-    }
-}

crates/dump/src/reader/v4/settings.rs

@@ -192,19 +192,14 @@ pub struct Facets {
     pub min_level_size: Option<NonZeroUsize>,
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Copy)]
+#[derive(Default, Debug, Clone, PartialEq, Eq, Copy)]
 pub enum Setting<T> {
     Set(T),
     Reset,
+    #[default]
     NotSet,
 }
 
-impl<T> Default for Setting<T> {
-    fn default() -> Self {
-        Self::NotSet
-    }
-}
-
 impl<T> Setting<T> {
     pub fn set(self) -> Option<T> {
         match self {

crates/dump/src/reader/v5/settings.rs

@@ -47,20 +47,15 @@ pub struct Settings<T> {
     pub _kind: PhantomData<T>,
 }
 
-#[derive(Debug, Clone, PartialEq, Eq, Copy)]
+#[derive(Default, Debug, Clone, PartialEq, Eq, Copy)]
 #[cfg_attr(test, derive(serde::Serialize))]
 pub enum Setting<T> {
     Set(T),
     Reset,
+    #[default]
     NotSet,
 }
 
-impl<T> Default for Setting<T> {
-    fn default() -> Self {
-        Self::NotSet
-    }
-}
-
 impl<T> Setting<T> {
     pub fn set(self) -> Option<T> {
         match self {

crates/dump/src/reader/v5/tasks.rs

@@ -322,7 +322,7 @@ impl From<Task> for TaskView {
             _ => None,
         });
 
-        let duration = finished_at.zip(started_at).map(|(tf, ts)| (tf - ts));
+        let duration = finished_at.zip(started_at).map(|(tf, ts)| tf - ts);
 
         Self {
             uid: id,

crates/dump/src/reader/v6/mod.rs

@@ -95,7 +95,8 @@ impl V6Reader {
             Err(e) => return Err(e.into()),
         };
 
-        let network = match fs::read(dump.path().join("network.json")) {
+        let mut network: Option<meilisearch_types::network::Network> =
+            match fs::read(dump.path().join("network.json")) {
                 Ok(network_file) => Some(serde_json::from_reader(&*network_file)?),
                 Err(error) => match error.kind() {
                     // Allows the file to be missing, this will only result in all experimental features disabled.
@@ -107,6 +108,14 @@ impl V6Reader {
                 },
             };
 
+        if let Some(network) = &mut network {
+            // as dumps are typically imported in a different machine as the emitter (otherwise dumpless upgrade would be used),
+            // we decide to remove the self to avoid alias issues
+            network.local = None;
+            // for the same reason we disable automatic sharding
+            network.leader = None;
+        }
+
         let webhooks = match fs::read(dump.path().join("webhooks.json")) {
             Ok(webhooks_file) => Some(serde_json::from_reader(&*webhooks_file)?),
             Err(error) => match error.kind() {

crates/fuzzers/src/bin/fuzz-indexing.rs

@@ -12,7 +12,7 @@ use milli::documents::mmap_from_objects;
 use milli::heed::EnvOpenOptions;
 use milli::progress::Progress;
 use milli::update::new::indexer;
-use milli::update::IndexerConfig;
+use milli::update::{IndexerConfig, MissingDocumentPolicy};
 use milli::vector::RuntimeEmbedders;
 use milli::Index;
 use serde_json::Value;
@@ -113,9 +113,12 @@ fn main() {
 
                             for op in &operations {
                                 match op {
-                                    Either::Left(documents) => {
+                                    Either::Left(documents) => indexer
-                                        indexer.replace_documents(documents).unwrap()
+                                        .replace_documents(
-                                    }
+                                            documents,
+                                            MissingDocumentPolicy::default(),
+                                        )
+                                        .unwrap(),
                                     Either::Right(ids) => indexer.delete_documents(ids),
                                 }
                             }

crates/index-scheduler/Cargo.toml

@@ -24,6 +24,7 @@ dump = { path = "../dump" }
 enum-iterator = "2.3.0"
 file-store = { path = "../file-store" }
 flate2 = "1.1.5"
+hashbrown = "0.15.5"
 indexmap = "2.12.0"
 meilisearch-auth = { path = "../meilisearch-auth" }
 meilisearch-types = { path = "../meilisearch-types" }
@@ -46,10 +47,14 @@ time = { version = "0.3.44", features = [
 tracing = "0.1.41"
 ureq = "2.12.1"
 uuid = { version = "1.18.1", features = ["serde", "v4"] }
-backoff = "0.4.0"
+backoff = { version = "0.4.0", features = ["futures", "tokio"] }
-reqwest = { version = "0.12.24", features = ["rustls-tls", "http2"], default-features = false }
+reqwest = { version = "0.12.24", features = [
+    "rustls-tls",
+    "http2",
+], default-features = false }
 rusty-s3 = "0.8.1"
 tokio = { version = "1.48.0", features = ["full"] }
+urlencoding = "2.1.3"
 
 [dev-dependencies]
 big_s = "1.0.2"
@@ -58,3 +63,6 @@ crossbeam-channel = "0.5.15"
 insta = { version = "=1.39.0", features = ["json", "redactions"] }
 maplit = "1.0.2"
 meili-snap = { path = "../meili-snap" }
+
+[features]
+enterprise = ["meilisearch-types/enterprise"]

crates/index-scheduler/src/dump.rs

@@ -3,17 +3,17 @@
 use std::collections::HashMap;
 use std::io;
 
+use crate::{utils, Error, IndexScheduler, Result};
 use dump::{KindDump, TaskDump, UpdateFile};
 use meilisearch_types::batches::{Batch, BatchId};
 use meilisearch_types::heed::RwTxn;
 use meilisearch_types::index_uid_pattern::IndexUidPattern;
 use meilisearch_types::milli;
+use meilisearch_types::milli::update::MissingDocumentPolicy;
 use meilisearch_types::tasks::{Kind, KindWithContent, Status, Task};
 use roaring::RoaringBitmap;
 use uuid::Uuid;
 
-use crate::{utils, Error, IndexScheduler, Result};
-
 pub struct Dump<'a> {
     index_scheduler: &'a IndexScheduler,
     wtxn: RwTxn<'a>,
@@ -164,6 +164,7 @@ impl<'a> Dump<'a> {
                     content_file: content_uuid.ok_or(Error::CorruptedDump)?,
                     documents_count,
                     allow_index_creation,
+                    on_missing_document: MissingDocumentPolicy::default(),
                 },
                 KindDump::DocumentDeletion { documents_ids } => KindWithContent::DocumentDeletion {
                     documents_ids,
@@ -238,6 +239,9 @@ impl<'a> Dump<'a> {
                 KindDump::IndexCompaction { index_uid } => {
                     KindWithContent::IndexCompaction { index_uid }
                 }
+                KindDump::NetworkTopologyChange(network_topology_change) => {
+                    KindWithContent::NetworkTopologyChange(network_topology_change)
+                }
             },
         };
 

crates/index-scheduler/src/error.rs

@@ -3,10 +3,13 @@ use std::fmt::Display;
 use meilisearch_types::batches::BatchId;
 use meilisearch_types::error::{Code, ErrorCode};
 use meilisearch_types::milli::index::RollbackOutcome;
+use meilisearch_types::milli::DocumentId;
+use meilisearch_types::tasks::network::ReceiveTaskError;
 use meilisearch_types::tasks::{Kind, Status};
 use meilisearch_types::{heed, milli};
 use reqwest::StatusCode;
 use thiserror::Error;
+use uuid::Uuid;
 
 use crate::TaskId;
 
@@ -191,6 +194,17 @@ pub enum Error {
     #[error(transparent)]
     HeedTransaction(heed::Error),
 
+    #[error("No network topology change task is currently enqueued or processing")]
+    ImportTaskWithoutNetworkTask,
+    #[error("The network task version (`{network_task}`) does not match the import task version (`{import_task}`)")]
+    NetworkVersionMismatch { network_task: Uuid, import_task: Uuid },
+    #[error("The import task emanates from an unknown remote `{0}`")]
+    ImportTaskUnknownRemote(String),
+    #[error("The import task with key `{0}` was already received")]
+    ImportTaskAlreadyReceived(DocumentId),
+    #[error("{action} requires the Enterprise Edition")]
+    RequiresEnterpriseEdition { action: &'static str },
+
     #[cfg(test)]
     #[error("Planned failure for tests.")]
     PlannedFailure,
@@ -248,6 +262,11 @@ impl Error {
             | Error::Persist(_)
             | Error::FeatureNotEnabled(_)
             | Error::Export(_)
+            | Error::ImportTaskWithoutNetworkTask
+            | Error::NetworkVersionMismatch { .. }
+            | Error::ImportTaskAlreadyReceived(_)
+            | Error::ImportTaskUnknownRemote(_)
+            | Error::RequiresEnterpriseEdition { .. }
             | Error::Anyhow(_) => true,
             Error::CreateBatch(_)
             | Error::CorruptedTaskQueue
@@ -307,6 +326,11 @@ impl ErrorCode for Error {
             Error::TaskDeletionWithEmptyQuery => Code::MissingTaskFilters,
             Error::TaskCancelationWithEmptyQuery => Code::MissingTaskFilters,
             Error::NoSpaceLeftInTaskQueue => Code::NoSpaceLeftOnDevice,
+            Error::ImportTaskWithoutNetworkTask => Code::ImportTaskWithoutNetworkTask,
+            Error::NetworkVersionMismatch { .. } => Code::NetworkVersionMismatch,
+            Error::ImportTaskAlreadyReceived(_) => Code::ImportTaskAlreadyReceived,
+            Error::ImportTaskUnknownRemote(_) => Code::ImportTaskUnknownRemote,
+            Error::RequiresEnterpriseEdition { .. } => Code::RequiresEnterpriseEdition,
             Error::S3Error { status, .. } if status.is_client_error() => {
                 Code::InvalidS3SnapshotRequest
             }
@@ -345,3 +369,12 @@ impl ErrorCode for Error {
         }
     }
 }
+
+impl From<ReceiveTaskError> for Error {
+    fn from(value: ReceiveTaskError) -> Self {
+        match value {
+            ReceiveTaskError::UnknownRemote(unknown) => Error::ImportTaskUnknownRemote(unknown),
+            ReceiveTaskError::DuplicateTask(dup) => Error::ImportTaskAlreadyReceived(dup),
+        }
+    }
+}

crates/index-scheduler/src/features.rs

@@ -38,6 +38,10 @@ impl RoFeatures {
         Self { runtime }
     }
 
+    pub fn from_runtime_features(features: RuntimeTogglableFeatures) -> Self {
+        Self { runtime: features }
+    }
+
     pub fn runtime_features(&self) -> RuntimeTogglableFeatures {
         self.runtime
     }

crates/index-scheduler/src/index_mapper/mod.rs

@@ -361,6 +361,12 @@ impl IndexMapper {
         Ok(())
     }
 
+    /// The number of indexes in the database
+    #[cfg(feature = "enterprise")] // only used in enterprise edition for now
+    pub fn index_count(&self, rtxn: &RoTxn) -> Result<u64> {
+        Ok(self.index_mapping.len(rtxn)?)
+    }
+
     /// Return an index, may open it if it wasn't already opened.
     pub fn index(&self, rtxn: &RoTxn, name: &str) -> Result<Index> {
         if let Some((current_name, current_index)) =

crates/index-scheduler/src/insta_snapshot.rs

@@ -6,7 +6,7 @@ use meilisearch_types::heed::types::{SerdeBincode, SerdeJson, Str};
 use meilisearch_types::heed::{Database, RoTxn};
 use meilisearch_types::milli::{CboRoaringBitmapCodec, RoaringBitmapCodec, BEU32};
 use meilisearch_types::tasks::{Details, Kind, Status, Task};
-use meilisearch_types::versioning;
+use meilisearch_types::versioning::{self, VERSION_MAJOR, VERSION_MINOR, VERSION_PATCH};
 use roaring::RoaringBitmap;
 
 use crate::index_mapper::IndexMapper;
@@ -27,6 +27,7 @@ pub fn snapshot_index_scheduler(scheduler: &IndexScheduler) -> String {
         queue,
         scheduler,
         persisted,
+        export_default_payload_size_bytes: _,
 
         index_mapper,
         features: _,
@@ -320,11 +321,18 @@ fn snapshot_details(d: &Details) -> String {
             format!("{{ url: {url:?}, api_key: {api_key:?}, payload_size: {payload_size:?}, indexes: {indexes:?} }}")
         }
         Details::UpgradeDatabase { from, to } => {
+            if to == &(VERSION_MAJOR, VERSION_MINOR, VERSION_PATCH) {
+                format!("{{ from: {from:?}, to: [current version] }}")
+            } else {
                 format!("{{ from: {from:?}, to: {to:?} }}")
             }
+        }
         Details::IndexCompaction { index_uid, pre_compaction_size, post_compaction_size } => {
             format!("{{ index_uid: {index_uid:?}, pre_compaction_size: {pre_compaction_size:?}, post_compaction_size: {post_compaction_size:?} }}")
         }
+        Details::NetworkTopologyChange { moved_documents, message } => {
+            format!("{{ moved_documents: {moved_documents:?}, message: {message:?}")
+        }
     }
 }
 
@@ -400,7 +408,21 @@ pub fn snapshot_batch(batch: &Batch) -> String {
 
     snap.push('{');
     snap.push_str(&format!("uid: {uid}, "));
-    snap.push_str(&format!("details: {}, ", serde_json::to_string(details).unwrap()));
+    let details = if let Some(upgrade_to) = &details.upgrade_to {
+        if upgrade_to.as_str()
+            == format!("v{VERSION_MAJOR}.{VERSION_MINOR}.{VERSION_PATCH}").as_str()
+        {
+            let mut details = details.clone();
+
+            details.upgrade_to = Some("[current version]".into());
+            serde_json::to_string(&details).unwrap()
+        } else {
+            serde_json::to_string(details).unwrap()
+        }
+    } else {
+        serde_json::to_string(details).unwrap()
+    };
+    snap.push_str(&format!("details: {details}, "));
     snap.push_str(&format!("stats: {}, ", serde_json::to_string(&stats).unwrap()));
     if !embedder_stats.skip_serializing() {
         snap.push_str(&format!(

crates/index-scheduler/src/lib.rs

@@ -48,6 +48,7 @@ use std::path::{Path, PathBuf};
 use std::sync::{Arc, RwLock};
 use std::time::Duration;
 
+use byte_unit::Byte;
 use dump::Dump;
 pub use error::Error;
 pub use features::RoFeatures;
@@ -68,10 +69,12 @@ use meilisearch_types::milli::vector::{
 use meilisearch_types::milli::{self, Index};
 use meilisearch_types::network::Network;
 use meilisearch_types::task_view::TaskView;
-use meilisearch_types::tasks::{KindWithContent, Task, TaskNetwork};
+use meilisearch_types::tasks::network::{
+    DbTaskNetwork, ImportData, ImportMetadata, Origin, TaskNetwork,
+};
+use meilisearch_types::tasks::{KindWithContent, Task};
 use meilisearch_types::webhooks::{Webhook, WebhooksDumpView, WebhooksView};
 use milli::vector::db::IndexEmbeddingConfig;
-use processing::ProcessingTasks;
 pub use queue::Query;
 use queue::Queue;
 use roaring::RoaringBitmap;
@@ -82,6 +85,7 @@ use uuid::Uuid;
 use versioning::Versioning;
 
 use crate::index_mapper::IndexMapper;
+use crate::processing::ProcessingTasks;
 use crate::utils::clamp_to_page_size;
 
 pub(crate) type BEI128 = I128<BE>;
@@ -144,9 +148,11 @@ pub struct IndexSchedulerOptions {
     /// If the autobatcher is allowed to automatically batch tasks
     /// it will only batch this defined maximum size (in bytes) of tasks at once.
     pub batched_tasks_size_limit: u64,
+    /// The maximum size of the default payload for exporting documents, in bytes
+    pub export_default_payload_size_bytes: Byte,
     /// The experimental features enabled for this instance.
     pub instance_features: InstanceTogglableFeatures,
-    /// The experimental features enabled for this instance.
+    /// Whether the index scheduler is able to auto upgrade or not.
     pub auto_upgrade: bool,
     /// The maximal number of entries in the search query cache of an embedder.
     ///
@@ -199,6 +205,9 @@ pub struct IndexScheduler {
     /// to the same embeddings for the same input text.
     embedders: Arc<RwLock<HashMap<EmbedderOptions, Arc<Embedder>>>>,
 
+    /// The maximum size of the default payload for exporting documents, in bytes
+    pub export_default_payload_size_bytes: Byte,
+
     // ================= test
     // The next entry is dedicated to the tests.
     /// Provide a way to set a breakpoint in multiple part of the scheduler.
@@ -234,6 +243,7 @@ impl IndexScheduler {
             cleanup_enabled: self.cleanup_enabled,
             experimental_no_edition_2024_for_dumps: self.experimental_no_edition_2024_for_dumps,
             persisted: self.persisted,
+            export_default_payload_size_bytes: self.export_default_payload_size_bytes,
 
             webhooks: self.webhooks.clone(),
             embedders: self.embedders.clone(),
@@ -345,6 +355,7 @@ impl IndexScheduler {
             persisted,
             webhooks: Arc::new(webhooks),
             embedders: Default::default(),
+            export_default_payload_size_bytes: options.export_default_payload_size_bytes,
 
             #[cfg(test)] // Will be replaced in `new_tests` in test environments
             test_breakpoint_sdr: crossbeam_channel::bounded(0).0,
@@ -700,14 +711,14 @@ impl IndexScheduler {
         self.queue.get_task_ids_from_authorized_indexes(&rtxn, query, filters, &processing)
     }
 
-    pub fn set_task_network(&self, task_id: TaskId, network: TaskNetwork) -> Result<()> {
+    pub fn set_task_network(&self, task_id: TaskId, network: DbTaskNetwork) -> Result<Task> {
         let mut wtxn = self.env.write_txn()?;
         let mut task =
             self.queue.tasks.get_task(&wtxn, task_id)?.ok_or(Error::TaskNotFound(task_id))?;
         task.network = Some(network);
         self.queue.tasks.all_tasks.put(&mut wtxn, &task_id, &task)?;
         wtxn.commit()?;
-        Ok(())
+        Ok(task)
     }
 
     /// Return the batches matching the query from the user's point of view along
@@ -757,18 +768,30 @@ impl IndexScheduler {
         task_id: Option<TaskId>,
         dry_run: bool,
     ) -> Result<Task> {
-        self.register_with_custom_metadata(kind, task_id, None, dry_run)
+        self.register_with_custom_metadata(kind, task_id, None, dry_run, None)
     }
 
     /// Register a new task in the scheduler, with metadata.
     ///
     /// If it fails and data was associated with the task, it tries to delete the associated data.
+    ///
+    /// # Parameters
+    ///
+    /// - task_network: network of the task to check.
+    ///
+    /// If the task is an import task, only accept it if:
+    ///
+    /// 1. There is an ongoing network topology change task
+    /// 2. The task to register matches the network version of the network topology change task
+    ///
+    /// Always accept the task if it is not an import task.
     pub fn register_with_custom_metadata(
         &self,
         kind: KindWithContent,
         task_id: Option<TaskId>,
         custom_metadata: Option<String>,
         dry_run: bool,
+        task_network: Option<TaskNetwork>,
     ) -> Result<Task> {
         // if the task doesn't delete or cancel anything and 40% of the task queue is full, we must refuse to enqueue the incoming task
         if !matches!(&kind, KindWithContent::TaskDeletion { tasks, .. } | KindWithContent::TaskCancelation { tasks, .. } if !tasks.is_empty())
@@ -779,7 +802,19 @@ impl IndexScheduler {
         }
 
         let mut wtxn = self.env.write_txn()?;
-        let task = self.queue.register(&mut wtxn, &kind, task_id, custom_metadata, dry_run)?;
+
+        if let Some(TaskNetwork::Import { import_from, network_change, metadata }) = &task_network {
+            self.update_network_task(&mut wtxn, import_from, network_change, metadata)?;
+        }
+
+        let task = self.queue.register(
+            &mut wtxn,
+            &kind,
+            task_id,
+            custom_metadata,
+            dry_run,
+            task_network.map(DbTaskNetwork::from),
+        )?;
 
         // If the registered task is a task cancelation
         // we inform the processing tasks to stop (if necessary).
@@ -801,6 +836,91 @@ impl IndexScheduler {
         Ok(task)
     }
 
+    pub fn network_no_index_for_remote(
+        &self,
+        remote_name: String,
+        origin: Origin,
+    ) -> Result<(), Error> {
+        let mut wtxn = self.env.write_txn()?;
+
+        self.update_network_task(
+            &mut wtxn,
+            &ImportData { remote_name, index_name: None, document_count: 0 },
+            &origin,
+            &ImportMetadata { index_count: 0, task_key: None, total_index_documents: 0 },
+        )?;
+
+        wtxn.commit()?;
+
+        // wake up the scheduler as the task state has changed
+        self.scheduler.wake_up.signal();
+
+        Ok(())
+    }
+
+    fn update_network_task(
+        &self,
+        wtxn: &mut heed::RwTxn<'_>,
+        import_from: &ImportData,
+        network_change: &Origin,
+        metadata: &ImportMetadata,
+    ) -> Result<(), Error> {
+        let mut network_tasks = self
+            .queue
+            .tasks
+            .get_kind(&*wtxn, meilisearch_types::tasks::Kind::NetworkTopologyChange)?;
+        if network_tasks.is_empty() {
+            return Err(Error::ImportTaskWithoutNetworkTask);
+        }
+        let network_task = {
+            let processing = self.processing_tasks.read().unwrap().processing.clone();
+            if processing.is_disjoint(&network_tasks) {
+                let enqueued = self
+                    .queue
+                    .tasks
+                    .get_status(&*wtxn, meilisearch_types::tasks::Status::Enqueued)?;
+
+                network_tasks &= enqueued;
+                if let Some(network_task) = network_tasks.into_iter().next() {
+                    network_task
+                } else {
+                    return Err(Error::ImportTaskWithoutNetworkTask);
+                }
+            } else {
+                network_tasks &= &*processing;
+                network_tasks.into_iter().next().unwrap()
+            }
+        };
+        let mut network_task = self.queue.tasks.get_task(&*wtxn, network_task)?.unwrap();
+        let network_task_version = network_task
+            .network
+            .as_ref()
+            .map(|network| network.network_version())
+            .unwrap_or_default();
+        if network_task_version != network_change.network_version {
+            return Err(Error::NetworkVersionMismatch {
+                network_task: network_task_version,
+                import_task: network_change.network_version,
+            });
+        }
+        let KindWithContent::NetworkTopologyChange(network_topology_change) =
+            &mut network_task.kind
+        else {
+            tracing::error!("unexpected network kind for network task while registering task");
+            return Err(Error::CorruptedTaskQueue);
+        };
+        network_topology_change.receive_remote_task(
+            &import_from.remote_name,
+            import_from.index_name.as_deref(),
+            metadata.task_key,
+            import_from.document_count,
+            metadata.index_count,
+            metadata.total_index_documents,
+        )?;
+        self.queue.tasks.update_task(wtxn, &mut network_task)?;
+        Ok(())
+    }
+
     /// Register a new task coming from a dump in the scheduler.
     /// By taking a mutable ref we're pretty sure no one will ever import a dump while actix is running.
     pub fn register_dumped_task(&mut self) -> Result<Dump<'_>> {

crates/index-scheduler/src/processing.rs

@@ -42,12 +42,10 @@ impl ProcessingTasks {
 
     /// Set the processing tasks to an empty list
     pub fn stop_processing(&mut self) -> Self {
-        self.progress = None;
-
         Self {
             batch: std::mem::take(&mut self.batch),
             processing: std::mem::take(&mut self.processing),
-            progress: None,
+            progress: std::mem::take(&mut self.progress),
         }
     }
 

crates/index-scheduler/src/queue/batches_test.rs

@@ -1,14 +1,17 @@
+use crate::insta_snapshot::{snapshot_bitmap, snapshot_index_scheduler};
+use crate::test_utils::Breakpoint::*;
+use crate::test_utils::{
+    index_creation_task, replace_document_import_task, replace_document_import_task_with_opts,
+    sample_documents, FailureLocation,
+};
+use crate::{IndexScheduler, Query};
 use meili_snap::snapshot;
 use meilisearch_auth::AuthFilter;
 use meilisearch_types::index_uid_pattern::IndexUidPattern;
+use meilisearch_types::milli::update::MissingDocumentPolicy;
 use meilisearch_types::tasks::{IndexSwap, KindWithContent, Status};
 use time::{Duration, OffsetDateTime};
 
-use crate::insta_snapshot::{snapshot_bitmap, snapshot_index_scheduler};
-use crate::test_utils::Breakpoint::*;
-use crate::test_utils::{index_creation_task, FailureLocation};
-use crate::{IndexScheduler, Query};
-
 #[test]
 fn query_batches_from_and_limit() {
     let (index_scheduler, mut handle) = IndexScheduler::test(true, vec![]);
@@ -487,3 +490,41 @@ fn query_batches_canceled_by() {
     // Return only 1 because the user is not authorized to see task 2
     snapshot!(snapshot_bitmap(&batches), @"[1,]");
 }
+
+#[test]
+fn batch_skip_creation_with_deletion() {
+    let (index_scheduler, mut handle) = IndexScheduler::test(true, vec![]);
+    let kind = index_creation_task("docs", "id");
+    let _task = index_scheduler.register(kind, None, false).unwrap();
+
+    handle.advance_one_successful_batch();
+
+    let (file0, documents_count0) = sample_documents(&index_scheduler, 1, 1);
+    let (file1, documents_count1) = sample_documents(&index_scheduler, 2, 1);
+    file0.persist().unwrap();
+    file1.persist().unwrap();
+    let kind = replace_document_import_task("docs", Some("id"), 1, documents_count0);
+    index_scheduler.register(kind, None, false).unwrap();
+    index_scheduler
+        .register(
+            KindWithContent::DocumentDeletion {
+                index_uid: "docs".to_string(),
+                documents_ids: vec!["1".to_string()],
+            },
+            None,
+            false,
+        )
+        .unwrap();
+    let kind = replace_document_import_task_with_opts(
+        "docs",
+        Some("id"),
+        2,
+        documents_count1,
+        MissingDocumentPolicy::Skip,
+    );
+    index_scheduler.register(kind, None, false).unwrap();
+
+    handle.advance_one_successful_batch();
+
+    snapshot!(snapshot_index_scheduler(&index_scheduler));
+}

crates/index-scheduler/src/queue/mod.rs

@@ -15,6 +15,7 @@ use file_store::FileStore;
 use meilisearch_types::batches::BatchId;
 use meilisearch_types::heed::{Database, Env, RoTxn, RwTxn, WithoutTls};
 use meilisearch_types::milli::{CboRoaringBitmapCodec, BEU32};
+use meilisearch_types::tasks::network::DbTaskNetwork;
 use meilisearch_types::tasks::{Kind, KindWithContent, Status, Task};
 use roaring::RoaringBitmap;
 use time::format_description::well_known::Rfc3339;
@@ -259,6 +260,7 @@ impl Queue {
         task_id: Option<TaskId>,
         custom_metadata: Option<String>,
         dry_run: bool,
+        network: Option<DbTaskNetwork>,
     ) -> Result<Task> {
         let next_task_id = self.tasks.next_task_id(wtxn)?;
 
@@ -280,7 +282,7 @@ impl Queue {
             details: kind.default_details(),
             status: Status::Enqueued,
             kind: kind.clone(),
-            network: None,
+            network,
             custom_metadata,
         };
         // For deletion and cancelation tasks, we want to make extra sure that they
@@ -348,6 +350,7 @@ impl Queue {
             None,
             None,
             false,
+            None,
         )?;
 
         Ok(())

crates/index-scheduler/src/queue/snapshots/batches_test.rs/batch_skip_creation_with_deletion/1.snap

@@ -0,0 +1,81 @@
+---
+source: crates/index-scheduler/src/queue/batches_test.rs
+---
+### Autobatching Enabled = true
+### Processing batch None:
+[]
+----------------------------------------------------------------------
+### All Tasks:
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { primary_key: Some("id"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "docs", primary_key: Some("id") }}
+1 {uid: 1, batch_uid: 1, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "docs", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
+2 {uid: 2, batch_uid: 1, status: succeeded, details: { received_document_ids: 1, deleted_documents: Some(1) }, kind: DocumentDeletion { index_uid: "docs", documents_ids: ["1"] }}
+3 {uid: 3, batch_uid: 1, status: succeeded, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "docs", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true, on_missing_document: Skip }}
+----------------------------------------------------------------------
+### Status:
+enqueued []
+succeeded [0,1,2,3,]
+----------------------------------------------------------------------
+### Kind:
+"documentAdditionOrUpdate" [1,3,]
+"documentDeletion" [2,]
+"indexCreation" [0,]
+----------------------------------------------------------------------
+### Index Tasks:
+docs [0,1,2,3,]
+----------------------------------------------------------------------
+### Index Mapper:
+docs: { number_of_documents: 0, field_distribution: {} }
+
+----------------------------------------------------------------------
+### Canceled By:
+
+----------------------------------------------------------------------
+### Enqueued At:
+[timestamp] [0,]
+[timestamp] [1,]
+[timestamp] [2,]
+[timestamp] [3,]
+----------------------------------------------------------------------
+### Started At:
+[timestamp] [0,]
+[timestamp] [1,2,3,]
+----------------------------------------------------------------------
+### Finished At:
+[timestamp] [0,]
+[timestamp] [1,2,3,]
+----------------------------------------------------------------------
+### All Batches:
+0 {uid: 0, details: {"primaryKey":"id"}, stats: {"totalNbTasks":1,"status":{"succeeded":1},"types":{"indexCreation":1},"indexUids":{"docs":1}}, stop reason: "created batch containing only task with id 0 of type `indexCreation` that cannot be batched with any other task.", }
+1 {uid: 1, details: {"receivedDocuments":2,"indexedDocuments":1,"providedIds":1,"deletedDocuments":1}, stats: {"totalNbTasks":3,"status":{"succeeded":3},"types":{"documentAdditionOrUpdate":2,"documentDeletion":1},"indexUids":{"docs":3}}, stop reason: "batched all enqueued tasks", }
+----------------------------------------------------------------------
+### Batch to tasks mapping:
+0 [0,]
+1 [1,2,3,]
+----------------------------------------------------------------------
+### Batches Status:
+succeeded [0,1,]
+----------------------------------------------------------------------
+### Batches Kind:
+"documentAdditionOrUpdate" [1,]
+"documentDeletion" [1,]
+"indexCreation" [0,]
+----------------------------------------------------------------------
+### Batches Index Tasks:
+docs [0,1,]
+----------------------------------------------------------------------
+### Batches Enqueued At:
+[timestamp] [0,]
+[timestamp] [1,]
+[timestamp] [1,]
+----------------------------------------------------------------------
+### Batches Started At:
+[timestamp] [0,]
+[timestamp] [1,]
+----------------------------------------------------------------------
+### Batches Finished At:
+[timestamp] [0,]
+[timestamp] [1,]
+----------------------------------------------------------------------
+### File Store:
+
+----------------------------------------------------------------------

crates/index-scheduler/src/queue/snapshots/test.rs/register/everything_is_successfully_registered.snap

@@ -7,9 +7,9 @@ source: crates/index-scheduler/src/queue/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: Some("mouse"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "catto", primary_key: Some("mouse") }}
-1 {uid: 1, status: enqueued, details: { received_documents: 12, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 12, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 12, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 12, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 50, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 50, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 50, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 50, allow_index_creation: true, on_missing_document: Create }}
-3 {uid: 3, status: enqueued, details: { received_documents: 5000, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 5000, allow_index_creation: true }}
+3 {uid: 3, status: enqueued, details: { received_documents: 5000, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 5000, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,1,2,3,]

crates/index-scheduler/src/queue/tasks.rs

@@ -3,7 +3,8 @@ use std::ops::{Bound, RangeBounds};
 use meilisearch_types::heed::types::{DecodeIgnore, SerdeBincode, SerdeJson, Str};
 use meilisearch_types::heed::{Database, Env, RoTxn, RwTxn, WithoutTls};
 use meilisearch_types::milli::{CboRoaringBitmapCodec, RoaringBitmapCodec, BEU32};
-use meilisearch_types::tasks::{Kind, Status, Task};
+use meilisearch_types::tasks::network::DbTaskNetwork;
+use meilisearch_types::tasks::{Kind, KindWithContent, Status, Task};
 use roaring::{MultiOps, RoaringBitmap};
 use time::OffsetDateTime;
 
@@ -114,14 +115,15 @@ impl TaskQueue {
     /// - CorruptedTaskQueue: The task doesn't exist in the database
     pub(crate) fn update_task(&self, wtxn: &mut RwTxn, task: &mut Task) -> Result<()> {
         let old_task = self.get_task(wtxn, task.uid)?.ok_or(Error::CorruptedTaskQueue)?;
-        let reprocessing = old_task.status != Status::Enqueued;
+        // network topology tasks may be processed multiple times.
+        let maybe_reprocessing = old_task.status != Status::Enqueued
+            || task.kind.as_kind() == Kind::NetworkTopologyChange;
 
-        debug_assert!(old_task != *task);
         debug_assert_eq!(old_task.uid, task.uid);
 
         // If we're processing a task that failed it may already contains a batch_uid
         debug_assert!(
-            reprocessing || (old_task.batch_uid.is_none() && task.batch_uid.is_some()),
+            maybe_reprocessing || (old_task.batch_uid.is_none() && task.batch_uid.is_some()),
             "\n==> old: {old_task:?}\n==> new: {task:?}"
         );
 
@@ -143,13 +145,24 @@ impl TaskQueue {
             })?;
         }
 
+        // Avoids rewriting part of the network topology change because of TOCTOU errors
+        if let (
+            KindWithContent::NetworkTopologyChange(old_state),
+            KindWithContent::NetworkTopologyChange(new_state),
+        ) = (old_task.kind, &mut task.kind)
+        {
+            new_state.merge(old_state);
+            // the state possibly just changed, rewrite the details
+            task.details = Some(new_state.to_details());
+        }
+
         assert_eq!(
             old_task.enqueued_at, task.enqueued_at,
             "Cannot update a task's enqueued_at time"
         );
         if old_task.started_at != task.started_at {
             assert!(
-                reprocessing || old_task.started_at.is_none(),
+                maybe_reprocessing || old_task.started_at.is_none(),
                 "Cannot update a task's started_at time"
             );
             if let Some(started_at) = old_task.started_at {
@@ -161,7 +174,7 @@ impl TaskQueue {
         }
         if old_task.finished_at != task.finished_at {
             assert!(
-                reprocessing || old_task.finished_at.is_none(),
+                maybe_reprocessing || old_task.finished_at.is_none(),
                 "Cannot update a task's finished_at time"
             );
             if let Some(finished_at) = old_task.finished_at {
@@ -175,7 +188,16 @@ impl TaskQueue {
         task.network = match (old_task.network, task.network.take()) {
             (None, None) => None,
             (None, Some(network)) | (Some(network), None) => Some(network),
-            (Some(_), Some(network)) => Some(network),
+            (Some(left), Some(right)) => Some(match (left, right) {
+                (
+                    DbTaskNetwork::Remotes { remote_tasks: mut left, network_version: _ },
+                    DbTaskNetwork::Remotes { remote_tasks: mut right, network_version },
+                ) => {
+                    left.append(&mut right);
+                    DbTaskNetwork::Remotes { remote_tasks: left, network_version }
+                }
+                (_, right) => right,
+            }),
         };
 
         self.all_tasks.put(wtxn, &task.uid, task)?;

crates/index-scheduler/src/queue/test.rs

@@ -203,26 +203,30 @@ fn test_disable_auto_deletion_of_tasks() {
         )
         .unwrap();
 
+    {
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]" }), name: "task_queue_is_full");
-    drop(rtxn);
+    }
-    drop(proc);
 
     // now we're above the max number of tasks
     // and if we try to advance in the tick function no new task deletion should be enqueued
     handle.advance_till([Start, BatchCreated]);
+    {
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]", ".**.original_filter" => "[filter]", ".**.query" => "[query]" }), name: "task_deletion_have_not_been_enqueued");
-    drop(rtxn);
+    }
-    drop(proc);
 }
 
 #[test]
@@ -267,59 +271,69 @@ fn test_auto_deletion_of_tasks() {
         )
         .unwrap();
 
+    {
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]" }), name: "task_queue_is_full");
-    drop(rtxn);
+    }
-    drop(proc);
 
+    {
         // now we're above the max number of tasks
         // and if we try to advance in the tick function a new task deletion should be enqueued
         handle.advance_till([Start, BatchCreated]);
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]", ".**.original_filter" => "[filter]", ".**.query" => "[query]" }), name: "task_deletion_have_been_enqueued");
-    drop(rtxn);
+    }
-    drop(proc);
 
+    {
         handle.advance_till([InsideProcessBatch, ProcessBatchSucceeded, AfterProcessing]);
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]", ".**.original_filter" => "[filter]", ".**.query" => "[query]" }), name: "task_deletion_have_been_processed");
-    drop(rtxn);
+    }
-    drop(proc);
 
     handle.advance_one_failed_batch();
     // a new task deletion has been enqueued
     handle.advance_one_successful_batch();
+    {
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]", ".**.original_filter" => "[filter]", ".**.query" => "[query]" }), name: "after_the_second_task_deletion");
-    drop(rtxn);
+    }
-    drop(proc);
 
     handle.advance_one_failed_batch();
     handle.advance_one_successful_batch();
+    {
         let rtxn = index_scheduler.env.read_txn().unwrap();
         let proc = index_scheduler.processing_tasks.read().unwrap();
-    let tasks =
+        let tasks = index_scheduler
-        index_scheduler.queue.get_task_ids(&rtxn, &Query { ..Default::default() }, &proc).unwrap();
+            .queue
+            .get_task_ids(&rtxn, &Query { ..Default::default() }, &proc)
+            .unwrap();
         let tasks = index_scheduler.queue.tasks.get_existing_tasks(&rtxn, tasks).unwrap();
         snapshot!(json_string!(tasks, { "[].enqueuedAt" => "[date]", "[].startedAt" => "[date]", "[].finishedAt" => "[date]", ".**.original_filter" => "[filter]", ".**.query" => "[query]" }), name: "everything_has_been_processed");
-    drop(rtxn);
+    }
-    drop(proc);
 }
 
 #[test]

crates/index-scheduler/src/scheduler/autobatcher.rs

@@ -74,6 +74,7 @@ impl From<KindWithContent> for AutobatchKind {
             | KindWithContent::DumpCreation { .. }
             | KindWithContent::Export { .. }
             | KindWithContent::UpgradeDatabase { .. }
+            | KindWithContent::NetworkTopologyChange(_)
             | KindWithContent::SnapshotCreation => {
                 panic!("The autobatcher should never be called with tasks with special priority or that don't apply to an index.")
             }

crates/index-scheduler/src/scheduler/autobatcher_test.rs

@@ -1,12 +1,12 @@
-use meilisearch_types::milli::update::IndexDocumentsMethod::{
-    self, ReplaceDocuments, UpdateDocuments,
-};
-use meilisearch_types::tasks::{BatchStopReason, IndexSwap, KindWithContent};
-use uuid::Uuid;
-
 use self::autobatcher::{autobatch, BatchKind};
 use super::*;
 use crate::TaskId;
+use meilisearch_types::milli::update::IndexDocumentsMethod::{
+    self, ReplaceDocuments, UpdateDocuments,
+};
+use meilisearch_types::milli::update::MissingDocumentPolicy;
+use meilisearch_types::tasks::{BatchStopReason, IndexSwap, KindWithContent};
+use uuid::Uuid;
 
 #[macro_export]
 macro_rules! debug_snapshot {
@@ -40,6 +40,7 @@ fn doc_imp(
         content_file: Uuid::new_v4(),
         documents_count: 0,
         allow_index_creation,
+        on_missing_document: MissingDocumentPolicy::default(),
     }
 }
 

crates/index-scheduler/src/scheduler/community_edition.rs

@@ -0,0 +1,37 @@
+use meilisearch_types::milli::progress::Progress;
+use meilisearch_types::tasks::Task;
+
+use super::create_batch::Batch;
+use crate::scheduler::process_batch::ProcessBatchInfo;
+use crate::utils::ProcessingBatch;
+use crate::{Error, IndexScheduler, Result};
+
+impl IndexScheduler {
+    pub(super) fn process_network_index_batch(
+        &self,
+        _network_task: Task,
+        _inner_batch: Box<Batch>,
+        _current_batch: &mut ProcessingBatch,
+        _progress: Progress,
+    ) -> Result<(Vec<Task>, ProcessBatchInfo)> {
+        Err(Error::RequiresEnterpriseEdition { action: "processing a network task" })
+    }
+
+    pub(super) fn process_network_ready(
+        &self,
+        _task: Task,
+        _progress: Progress,
+    ) -> Result<(Vec<Task>, ProcessBatchInfo)> {
+        Err(Error::RequiresEnterpriseEdition { action: "processing a network task" })
+    }
+
+    #[cfg(unix)]
+    pub(super) async fn process_snapshot_to_s3(
+        &self,
+        _progress: Progress,
+        _opts: meilisearch_types::milli::update::S3SnapshotOptions,
+        _tasks: Vec<Task>,
+    ) -> Result<Vec<Task>> {
+        Err(Error::RequiresEnterpriseEdition { action: "processing an S3-streaming snapshot task" })
+    }
+}

crates/index-scheduler/src/scheduler/create_batch.rs

@@ -2,8 +2,9 @@ use std::fmt;
 use std::io::ErrorKind;
 
 use meilisearch_types::heed::RoTxn;
-use meilisearch_types::milli::update::IndexDocumentsMethod;
+use meilisearch_types::milli::update::{IndexDocumentsMethod, MissingDocumentPolicy};
 use meilisearch_types::settings::{Settings, Unchecked};
+use meilisearch_types::tasks::network::NetworkTopologyState;
 use meilisearch_types::tasks::{BatchStopReason, Kind, KindWithContent, Status, Task};
 use roaring::RoaringBitmap;
 use uuid::Uuid;
@@ -59,12 +60,20 @@ pub(crate) enum Batch {
         index_uid: String,
         task: Task,
     },
+    #[allow(clippy::enum_variant_names)] // warranted because we are executing an inner index batch
+    NetworkIndexBatch {
+        network_task: Task,
+        inner_batch: Box<Batch>,
+    },
+    NetworkReady {
+        task: Task,
+    },
 }
 
 #[derive(Debug)]
 pub(crate) enum DocumentOperation {
-    Replace(Uuid),
+    Replace { content_file: Uuid, on_missing_document: MissingDocumentPolicy },
-    Update(Uuid),
+    Update { content_file: Uuid, on_missing_document: MissingDocumentPolicy },
     Delete(Vec<String>),
 }
 
@@ -140,9 +149,14 @@ impl Batch {
                     ..
                 } => RoaringBitmap::from_iter(tasks.iter().chain(other).map(|task| task.uid)),
             },
-            Batch::IndexSwap { task } => {
+            Batch::IndexSwap { task } | Batch::NetworkReady { task } => {
                 RoaringBitmap::from_sorted_iter(std::iter::once(task.uid)).unwrap()
             }
+            Batch::NetworkIndexBatch { network_task, inner_batch } => {
+                let mut tasks = inner_batch.ids();
+                tasks.insert(network_task.uid);
+                tasks
+            }
         }
     }
 
@@ -156,12 +170,14 @@ impl Batch {
             | Dump(_)
             | Export { .. }
             | UpgradeDatabase { .. }
+            | NetworkReady { .. }
             | IndexSwap { .. } => None,
             IndexOperation { op, .. } => Some(op.index_uid()),
             IndexCreation { index_uid, .. }
             | IndexUpdate { index_uid, .. }
             | IndexDeletion { index_uid, .. }
             | IndexCompaction { index_uid, .. } => Some(index_uid),
+            NetworkIndexBatch { network_task: _, inner_batch } => inner_batch.index_uid(),
         }
     }
 }
@@ -184,6 +200,8 @@ impl fmt::Display for Batch {
             Batch::IndexCompaction { .. } => f.write_str("IndexCompaction")?,
             Batch::Export { .. } => f.write_str("Export")?,
             Batch::UpgradeDatabase { .. } => f.write_str("UpgradeDatabase")?,
+            Batch::NetworkIndexBatch { .. } => f.write_str("NetworkTopologyChange")?,
+            Batch::NetworkReady { .. } => f.write_str("NetworkTopologyChange")?,
         };
         match index_uid {
             Some(name) => f.write_fmt(format_args!(" on {name:?} from tasks: {tasks:?}")),
@@ -293,13 +311,22 @@ impl IndexScheduler {
                 for task in tasks.iter() {
                     match task.kind {
                         KindWithContent::DocumentAdditionOrUpdate {
-                            content_file, method, ..
+                            content_file,
+                            method,
+                            on_missing_document,
+                            ..
                         } => match method {
                             IndexDocumentsMethod::ReplaceDocuments => {
-                                operations.push(DocumentOperation::Replace(content_file))
+                                operations.push(DocumentOperation::Replace {
+                                    content_file,
+                                    on_missing_document,
+                                })
                             }
                             IndexDocumentsMethod::UpdateDocuments => {
-                                operations.push(DocumentOperation::Update(content_file))
+                                operations.push(DocumentOperation::Update {
+                                    content_file,
+                                    on_missing_document,
+                                })
                             }
                             _ => unreachable!("Unknown document merging method"),
                         },
@@ -452,6 +479,7 @@ impl IndexScheduler {
     pub(crate) fn create_next_batch(
         &self,
         rtxn: &RoTxn,
+        processing_network_tasks: &RoaringBitmap,
     ) -> Result<Option<(Batch, ProcessingBatch)>> {
         #[cfg(test)]
         self.maybe_fail(crate::test_utils::FailureLocation::InsideCreateBatch)?;
@@ -460,7 +488,6 @@ impl IndexScheduler {
         let mut current_batch = ProcessingBatch::new(batch_id);
 
         let enqueued = &self.queue.tasks.get_status(rtxn, Status::Enqueued)?;
-        let count_total_enqueued = enqueued.len();
         let failed = &self.queue.tasks.get_status(rtxn, Status::Failed)?;
 
         // 0. we get the last task to cancel.
@@ -509,7 +536,15 @@ impl IndexScheduler {
             )));
         }
 
-        // 2. we get the next task to delete
+        // 2. Check for enqueued network topology changes
+        let network_changes = self.queue.tasks.get_kind(rtxn, Kind::NetworkTopologyChange)?
+            & (enqueued | processing_network_tasks);
+        if let Some(task_id) = network_changes.iter().next() {
+            let task = self.queue.tasks.get_task(rtxn, task_id)?.unwrap();
+            return self.start_processing_network(rtxn, task, enqueued, current_batch);
+        }
+
+        // 3. we get the next task to delete
         let to_delete = self.queue.tasks.get_kind(rtxn, Kind::TaskDeletion)? & enqueued;
         if !to_delete.is_empty() {
             let mut tasks = self.queue.tasks.get_existing_tasks(rtxn, to_delete)?;
@@ -519,7 +554,7 @@ impl IndexScheduler {
             return Ok(Some((Batch::TaskDeletions(tasks), current_batch)));
         }
 
-        // 3. we get the next task to compact
+        // 4. we get the next task to compact
         let to_compact = self.queue.tasks.get_kind(rtxn, Kind::IndexCompaction)? & enqueued;
         if let Some(task_id) = to_compact.min() {
             let mut task =
@@ -534,7 +569,7 @@ impl IndexScheduler {
             return Ok(Some((Batch::IndexCompaction { index_uid, task }, current_batch)));
         }
 
-        // 4. we batch the export.
+        // 5. we batch the export.
         let to_export = self.queue.tasks.get_kind(rtxn, Kind::Export)? & enqueued;
         if !to_export.is_empty() {
             let task_id = to_export.iter().next().expect("There must be at least one export task");
@@ -545,7 +580,7 @@ impl IndexScheduler {
             return Ok(Some((Batch::Export { task }, current_batch)));
         }
 
-        // 5. we batch the snapshot.
+        // 6. we batch the snapshot.
         let to_snapshot = self.queue.tasks.get_kind(rtxn, Kind::SnapshotCreation)? & enqueued;
         if !to_snapshot.is_empty() {
             let mut tasks = self.queue.tasks.get_existing_tasks(rtxn, to_snapshot)?;
@@ -555,7 +590,7 @@ impl IndexScheduler {
             return Ok(Some((Batch::SnapshotCreation(tasks), current_batch)));
         }
 
-        // 6. we batch the dumps.
+        // 7. we batch the dumps.
         let to_dump = self.queue.tasks.get_kind(rtxn, Kind::DumpCreation)? & enqueued;
         if let Some(to_dump) = to_dump.min() {
             let mut task =
@@ -568,25 +603,66 @@ impl IndexScheduler {
             return Ok(Some((Batch::Dump(task), current_batch)));
         }
 
-        // 7. We make a batch from the unprioritised tasks. Start by taking the next enqueued task.
+        let network = self.network();
-        let task_id = if let Some(task_id) = enqueued.min() { task_id } else { return Ok(None) };
-        let mut task =
-            self.queue.tasks.get_task(rtxn, task_id)?.ok_or(Error::CorruptedTaskQueue)?;
 
+        // 8. We make a batch from the unprioritised tasks.
+        let (batch, current_batch) =
+            self.create_next_batch_unprioritized(rtxn, enqueued, current_batch, |task| {
+                // We want to execute all tasks, except those that have a version strictly higher than the network version
+
+                let Some(task_version) =
+                    task.network.as_ref().map(|tastk_network| tastk_network.network_version())
+                else {
+                    // do not skip tasks that have no network version, otherwise we will never execute them
+                    return false;
+                };
+
+                // skip tasks with a version strictly higher than the network version
+                task_version > network.version
+            })?;
+        Ok(batch.map(|batch| (batch, current_batch)))
+    }
+
+    fn create_next_batch_unprioritized<F>(
+        &self,
+        rtxn: &RoTxn,
+        enqueued: &RoaringBitmap,
+        mut current_batch: ProcessingBatch,
+        mut skip_if: F,
+    ) -> Result<(Option<Batch>, ProcessingBatch)>
+    where
+        F: FnMut(&Task) -> bool,
+    {
+        let count_total_enqueued = enqueued.len();
+
+        let mut enqueued_it = enqueued.iter();
+        let mut task;
+        let index_name = loop {
+            let Some(task_id) = enqueued_it.next() else {
+                return Ok((None, current_batch));
+            };
+            task = self.queue.tasks.get_task(rtxn, task_id)?.ok_or(Error::CorruptedTaskQueue)?;
+
+            if skip_if(&task) {
+                continue;
+            }
             // If the task is not associated with any index, verify that it is an index swap and
             // create the batch directly. Otherwise, get the index name associated with the task
             // and use the autobatcher to batch the enqueued tasks associated with it
 
-        let index_name = if let Some(&index_name) = task.indexes().first() {
+            if let Some(&index_name) = task.indexes().first() {
-            index_name
+                break index_name;
             } else {
-            assert!(matches!(&task.kind, KindWithContent::IndexSwap { swaps } if swaps.is_empty()));
+                assert!(
+                    matches!(&task.kind, KindWithContent::IndexSwap { swaps } if swaps.is_empty())
+                );
                 current_batch.processing(Some(&mut task));
                 current_batch.reason(BatchStopReason::TaskCannotBeBatched {
                     kind: Kind::IndexSwap,
                     id: task.uid,
                 });
-            return Ok(Some((Batch::IndexSwap { task }, current_batch)));
+                return Ok((Some(Batch::IndexSwap { task }), current_batch));
+            };
         };
 
         let index_already_exists = self.index_mapper.exists(rtxn, index_name)?;
@@ -621,6 +697,10 @@ impl IndexScheduler {
                 .get_task(rtxn, task_id)
                 .and_then(|task| task.ok_or(Error::CorruptedTaskQueue))?;
 
+            if skip_if(&task) {
+                continue;
+            }
+
             if let Some(uuid) = task.content_uuid() {
                 let content_size = match self.queue.file_store.compute_size(uuid) {
                     Ok(content_size) => content_size,
@@ -651,19 +731,127 @@ impl IndexScheduler {
             autobatcher::autobatch(enqueued, index_already_exists, primary_key.as_deref())
         {
             current_batch.reason(autobatch_stop_reason.unwrap_or(stop_reason));
-            return Ok(self
+            let batch = self.create_next_batch_index(
-                .create_next_batch_index(
                 rtxn,
                 index_name.to_string(),
                 batchkind,
                 &mut current_batch,
                 create_index,
-                )?
+            )?;
-                .map(|batch| (batch, current_batch)));
+            return Ok((batch, current_batch));
         }
 
         // If we found no tasks then we were notified for something that got autobatched
         // somehow and there is nothing to do.
-        Ok(None)
+        Ok((None, current_batch))
+    }
+
+    fn start_processing_network(
+        &self,
+        rtxn: &RoTxn,
+        mut task: Task,
+        enqueued: &RoaringBitmap,
+        mut current_batch: ProcessingBatch,
+    ) -> Result<Option<(Batch, ProcessingBatch)>> {
+        current_batch.processing(Some(&mut task));
+        current_batch.reason(BatchStopReason::NetworkTask { id: task.uid });
+
+        let change_version =
+            task.network.as_ref().map(|network| network.network_version()).unwrap_or_default();
+        let KindWithContent::NetworkTopologyChange(network_topology_change) = &task.kind else {
+            panic!("inconsistent kind with content")
+        };
+
+        match network_topology_change.state() {
+            NetworkTopologyState::WaitingForOlderTasks => {
+                let res =
+                    self.create_next_batch_unprioritized(rtxn, enqueued, current_batch, |task| {
+                        // in this limited mode of execution, we only want to run tasks:
+                        // 0. with an index
+                        // 1. with a version
+                        // 2. that version strictly lower than the network task version
+
+                        // 0. skip indexless tasks that are not index swap
+                        if task.index_uid().is_none() && task.kind.as_kind() != Kind::IndexSwap {
+                            return true;
+                        }
+
+                        // 1. skip tasks without version
+                        let Some(task_version) =
+                            task.network.as_ref().map(|network| network.network_version())
+                        else {
+                            return true;
+                        };
+
+                        // 2. skip tasks with a version equal or higher to the network task version
+                        task_version >= change_version
+                    });
+
+                let (batch, mut current_batch) = res?;
+
+                let batch = match batch {
+                    Some(batch) => {
+                        let inner_batch = Box::new(batch);
+                        let inner_reason = current_batch.reason.to_string();
+                        current_batch.reason(BatchStopReason::NetworkTaskOlderTasks {
+                            id: task.uid,
+                            inner_reason,
+                        });
+
+                        Batch::NetworkIndexBatch { network_task: task, inner_batch }
+                    }
+                    None => Batch::NetworkReady { task },
+                };
+
+                Ok(Some((batch, current_batch)))
+            }
+            NetworkTopologyState::ImportingDocuments => {
+                // if the import is done we need to go to the next state
+                if network_topology_change.is_import_finished() {
+                    return Ok(Some((Batch::NetworkReady { task }, current_batch)));
+                }
+
+                let res =
+                    self.create_next_batch_unprioritized(rtxn, enqueued, current_batch, |task| {
+                        // in this limited mode of execution, we only want to run tasks:
+                        // 0. with an index
+                        // 1. with a version
+                        // 2. that version equal to the network task version
+
+                        // 0. skip indexless tasks
+                        if task.index_uid().is_none() && task.kind.as_kind() != Kind::IndexSwap {
+                            return true;
+                        }
+
+                        // 1. skip tasks without version
+                        let Some(task_version) =
+                            task.network.as_ref().map(|network| network.network_version())
+                        else {
+                            return true;
+                        };
+
+                        // 2. skip tasks with a version different from the network task version
+                        task_version != change_version
+                    });
+
+                let (batch, mut current_batch) = res?;
+
+                let batch = batch.map(|batch| {
+                    let inner_batch = Box::new(batch);
+                    let inner_reason = current_batch.reason.to_string();
+                    current_batch.reason(BatchStopReason::NetworkTaskImportTasks {
+                        id: task.uid,
+                        inner_reason,
+                    });
+
+                    (Batch::NetworkIndexBatch { network_task: task, inner_batch }, current_batch)
+                });
+
+                Ok(batch)
+            }
+            NetworkTopologyState::ExportingDocuments | NetworkTopologyState::Finished => {
+                Ok(Some((Batch::NetworkReady { task }, current_batch)))
+            }
+        }
     }
 }

crates/index-scheduler/src/scheduler/enterprise_edition/mod.rs

@@ -0,0 +1,845 @@
+// Copyright © 2025 Meilisearch Some Rights Reserved
+// This file is part of Meilisearch Enterprise Edition (EE).
+// Use of this source code is governed by the Business Source License 1.1,
+// as found in the LICENSE-EE file or at <https://mariadb.com/bsl11>
+
+use std::time::Duration;
+
+use bumpalo::Bump;
+use roaring::RoaringBitmap;
+
+use meilisearch_types::milli::documents::PrimaryKey;
+use meilisearch_types::milli::progress::{EmbedderStats, Progress};
+use meilisearch_types::milli::update::new::indexer;
+use meilisearch_types::milli::update::new::indexer::current_edition::sharding::Shards;
+use meilisearch_types::milli::{self};
+use meilisearch_types::network::Remote;
+use meilisearch_types::tasks::network::{NetworkTopologyState, Origin};
+use meilisearch_types::tasks::{KindWithContent, Status, Task};
+
+use super::create_batch::Batch;
+use crate::scheduler::process_batch::ProcessBatchInfo;
+use crate::scheduler::process_export::{ExportContext, ExportOptions, TargetInstance};
+use crate::utils::ProcessingBatch;
+use crate::{Error, IndexScheduler, Result};
+
+impl IndexScheduler {
+    pub(super) fn process_network_index_batch(
+        &self,
+        mut network_task: Task,
+        inner_batch: Box<Batch>,
+        current_batch: &mut ProcessingBatch,
+        progress: Progress,
+    ) -> Result<(Vec<Task>, ProcessBatchInfo)> {
+        let KindWithContent::NetworkTopologyChange(network_topology_change) =
+            &mut network_task.kind
+        else {
+            tracing::error!("unexpected network kind for network task while processing batch");
+            return Err(Error::CorruptedTaskQueue);
+        };
+
+        let network = network_topology_change.network_for_state();
+
+        let (mut tasks, info) =
+            self.process_batch(*inner_batch, current_batch, progress, network)?;
+
+        for task in &tasks {
+            let Some(network) = task.network.as_ref() else {
+                continue;
+            };
+            let Some(import) = network.import_data() else {
+                continue;
+            };
+            if let Some(index_name) = import.index_name.as_deref() {
+                network_topology_change.process_remote_tasks(
+                    &import.remote_name,
+                    index_name,
+                    import.document_count,
+                );
+            }
+        }
+        network_task.details = Some(network_topology_change.to_details());
+
+        tasks.push(network_task);
+        Ok((tasks, info))
+    }
+
+    pub(super) fn process_network_ready(
+        &self,
+        mut task: Task,
+        progress: Progress,
+    ) -> Result<(Vec<Task>, ProcessBatchInfo)> {
+        let KindWithContent::NetworkTopologyChange(network_topology_change) = &mut task.kind else {
+            tracing::error!("network topology change task has the wrong kind with content");
+            return Err(Error::CorruptedTaskQueue);
+        };
+
+        let Some(task_network) = &task.network else {
+            tracing::error!("network topology change task has no network");
+            return Err(Error::CorruptedTaskQueue);
+        };
+
+        let origin;
+        let origin = match task_network.origin() {
+            Some(origin) => origin,
+            None => {
+                let myself = network_topology_change.in_name().expect("origin is not the leader");
+                origin = Origin {
+                    remote_name: myself.to_string(),
+                    task_uid: task.uid,
+                    network_version: task_network.network_version(),
+                };
+                &origin
+            }
+        };
+
+        let mut moved_documents = None;
+        if let (Some((remotes, out_name)), Some(new_shards)) =
+            (network_topology_change.export_to_process(), network_topology_change.new_shards())
+        {
+            moved_documents = Some(self.balance_documents(
+                remotes,
+                out_name,
+                new_shards,
+                origin,
+                &progress,
+                &self.scheduler.must_stop_processing,
+            )?);
+        }
+        if let Some(moved_documents) = moved_documents {
+            // we need the mut moved documents to avoid a lifetime error in the previous if let.
+            network_topology_change.set_moved(moved_documents);
+        }
+        network_topology_change.update_state();
+        if network_topology_change.state() == NetworkTopologyState::Finished {
+            task.status = Status::Succeeded;
+        }
+
+        task.details = Some(network_topology_change.to_details());
+        Ok((vec![task], Default::default()))
+    }
+
+    fn balance_documents<'a, I: Iterator<Item = (&'a str, &'a Remote)> + Clone>(
+        &self,
+        remotes: I,
+        out_name: &str,
+        new_shards: Shards,
+        network_change_origin: &Origin,
+        progress: &Progress,
+        must_stop_processing: &crate::scheduler::MustStopProcessing,
+    ) -> crate::Result<u64> {
+        // TECHDEBT: this spawns a `ureq` agent additionally to `reqwest`. We probably want to harmonize all of this.
+        let agent = ureq::AgentBuilder::new().timeout(Duration::from_secs(5)).build();
+
+        let mut indexer_alloc = Bump::new();
+
+        let scheduler_rtxn = self.env.read_txn()?;
+
+        let index_count = self.index_mapper.index_count(&scheduler_rtxn)?;
+
+        // when the instance is empty, we still need to tell that to remotes, as they cannot know of that fact and will be waiting for
+        // data
+        if index_count == 0 {
+            for (remote_name, remote) in remotes {
+                let target = TargetInstance {
+                    remote_name: Some(remote_name),
+                    base_url: &remote.url,
+                    api_key: remote.write_api_key.as_deref(),
+                };
+
+                let res = self.export_no_index(
+                    target,
+                    out_name,
+                    network_change_origin,
+                    &agent,
+                    must_stop_processing,
+                );
+
+                if let Err(err) = res {
+                    tracing::warn!("Could not signal not to wait documents to `{remote_name}` due to error: {err}");
+                }
+            }
+            return Ok(0);
+        }
+
+        let mut total_moved_documents = 0;
+
+        self.index_mapper.try_for_each_index::<(), ()>(
+            &scheduler_rtxn,
+            |index_uid, index| -> crate::Result<()> {
+                indexer_alloc.reset();
+                let err = |err| Error::from_milli(err, Some(index_uid.to_string()));
+                let index_rtxn = index.read_txn()?;
+                let all_docids = index.external_documents_ids();
+                let mut documents_to_move_to =
+                    hashbrown::HashMap::<String, RoaringBitmap>::new();
+                let mut documents_to_delete = RoaringBitmap::new();
+
+                for res in all_docids.iter(&index_rtxn)? {
+                    let (external_docid, docid) = res?;
+                    match new_shards.processing_shard(external_docid) {
+                        Some(shard) if shard.is_own => continue,
+                        Some(shard) => {
+                            documents_to_move_to.entry_ref(&shard.name).or_default().insert(docid);
+                        }
+                        None => {
+                            documents_to_delete.insert(docid);
+                        }
+                    }
+                }
+
+                let fields_ids_map = index.fields_ids_map(&index_rtxn)?;
+
+                for (remote_name, remote) in remotes.clone() {
+                    let documents_to_move =
+                        documents_to_move_to.remove(remote_name).unwrap_or_default();
+
+                    let target = TargetInstance {
+                        remote_name: Some(remote_name),
+                        base_url: &remote.url,
+                        api_key: remote.write_api_key.as_deref(),
+                    };
+                    let options = ExportOptions {
+                        index_uid,
+                        payload_size: None,
+                        override_settings: false,
+                        export_mode: super::process_export::ExportMode::NetworkBalancing {
+                            index_count,
+                            export_old_remote_name: out_name,
+                            network_change_origin,
+                        },
+                    };
+                    let ctx = ExportContext {
+                        index,
+                        index_rtxn: &index_rtxn,
+                        universe: &documents_to_move,
+                        progress,
+                        agent: &agent,
+                        must_stop_processing,
+                    };
+
+                    let res = self.export_one_index(target, options, ctx);
+
+                    match res {
+                        Ok(_) =>{ documents_to_delete |= documents_to_move;}
+                        Err(err) => {
+                            tracing::warn!("Could not export documents to `{remote_name}` due to error: {err}\n  - Note: Documents will be kept");
+                        }
+                    }
+
+
+                }
+
+                if documents_to_delete.is_empty() {
+                    return Ok(());
+                }
+
+                total_moved_documents += documents_to_delete.len();
+
+                self.delete_documents_from_index(progress, must_stop_processing, &indexer_alloc, index_uid, index, &err, index_rtxn, documents_to_delete, fields_ids_map)
+            },
+        )?;
+
+        Ok(total_moved_documents)
+    }
+
+    #[allow(clippy::too_many_arguments)]
+    fn delete_documents_from_index(
+        &self,
+        progress: &Progress,
+        must_stop_processing: &super::MustStopProcessing,
+        indexer_alloc: &Bump,
+        index_uid: &str,
+        index: &milli::Index,
+        err: &impl Fn(milli::Error) -> Error,
+        index_rtxn: milli::heed::RoTxn<'_, milli::heed::WithoutTls>,
+        documents_to_delete: RoaringBitmap,
+        fields_ids_map: milli::FieldsIdsMap,
+    ) -> std::result::Result<(), Error> {
+        let mut new_fields_ids_map = fields_ids_map.clone();
+
+        // candidates not empty => index not empty => a primary key is set
+        let primary_key = index.primary_key(&index_rtxn)?.unwrap();
+
+        let primary_key = PrimaryKey::new_or_insert(primary_key, &mut new_fields_ids_map)
+            .map_err(milli::Error::from)
+            .map_err(err)?;
+
+        let mut index_wtxn = index.write_txn()?;
+
+        let mut indexer = indexer::DocumentDeletion::new();
+        indexer.delete_documents_by_docids(documents_to_delete);
+        let document_changes = indexer.into_changes(indexer_alloc, primary_key);
+        let embedders = index
+            .embedding_configs()
+            .embedding_configs(&index_wtxn)
+            .map_err(milli::Error::from)
+            .map_err(err)?;
+        let embedders = self.embedders(index_uid.to_string(), embedders)?;
+        let indexer_config = self.index_mapper.indexer_config();
+        let pool = &indexer_config.thread_pool;
+
+        indexer::index(
+            &mut index_wtxn,
+            index,
+            pool,
+            indexer_config.grenad_parameters(),
+            &fields_ids_map,
+            new_fields_ids_map,
+            None, // document deletion never changes primary key
+            &document_changes,
+            embedders,
+            &|| must_stop_processing.get(),
+            progress,
+            &EmbedderStats::default(),
+        )
+        .map_err(err)?;
+
+        // update stats
+        let mut mapper_wtxn = self.env.write_txn()?;
+        let stats = crate::index_mapper::IndexStats::new(index, &index_wtxn).map_err(err)?;
+        self.index_mapper.store_stats_of(&mut mapper_wtxn, index_uid, &stats)?;
+
+        index_wtxn.commit()?;
+        // update stats after committing changes to index
+        mapper_wtxn.commit()?;
+
+        Ok(())
+    }
+
+    #[cfg(unix)]
+    async fn assume_role_with_web_identity(
+        role_arn: &str,
+        web_identity_token_file: &std::path::Path,
+    ) -> anyhow::Result<StsCredentials> {
+        use std::env::VarError;
+
+        let token = tokio::fs::read_to_string(web_identity_token_file)
+            .await
+            .map_err(|e| anyhow::anyhow!("Failed to read web identity token file: {e}"))?;
+
+        let duration: u32 =
+            match std::env::var("MEILI_EXPERIMENTAL_S3_WEB_IDENTITY_TOKEN_DURATION_SECONDS") {
+                Ok(s) => s.parse()?,
+                Err(VarError::NotPresent) => 3600,
+                Err(VarError::NotUnicode(e)) => {
+                    anyhow::bail!("Invalid duration: {e:?}")
+                }
+            };
+
+        let form_data = [
+            ("Action", "AssumeRoleWithWebIdentity"),
+            ("Version", "2011-06-15"),
+            ("RoleArn", role_arn),
+            ("RoleSessionName", "meilisearch-snapshot-session"),
+            ("WebIdentityToken", &token),
+            ("DurationSeconds", &duration.to_string()),
+        ];
+
+        let client = reqwest::Client::new();
+        let response = client
+            .post("https://sts.amazonaws.com/")
+            .header(reqwest::header::ACCEPT, "application/json")
+            .header(reqwest::header::CONTENT_TYPE, "application/x-www-form-urlencoded")
+            .form(&form_data)
+            .send()
+            .await
+            .map_err(|e| anyhow::anyhow!("Failed to send STS request: {e}"))?;
+
+        let status = response.status();
+        let body = response
+            .text()
+            .await
+            .map_err(|e| anyhow::anyhow!("Failed to read STS response body: {e}"))?;
+
+        if !status.is_success() {
+            return Err(anyhow::anyhow!("STS request failed with status {status}: {body}"));
+        }
+
+        let sts_response: StsResponse = serde_json::from_str(&body)
+            .map_err(|e| anyhow::anyhow!("Failed to deserialize STS response: {e}"))?;
+
+        Ok(sts_response.response.result.credentials)
+    }
+
+    #[cfg(unix)]
+    async fn extract_credentials_from_options(
+        s3_access_key: Option<String>,
+        s3_secret_key: Option<String>,
+        s3_role_arn: Option<String>,
+        s3_web_identity_token_file: Option<std::path::PathBuf>,
+    ) -> anyhow::Result<(String, String, Option<String>)> {
+        let static_credentials = s3_access_key.zip(s3_secret_key);
+        let web_identity = s3_role_arn.zip(s3_web_identity_token_file);
+        match (static_credentials, web_identity) {
+            (Some((access_key, secret_key)), None) => Ok((access_key, secret_key, None)),
+            (None, Some((role_arn, token_file))) => {
+                let StsCredentials { access_key_id, secret_access_key, session_token } =
+                    Self::assume_role_with_web_identity(&role_arn, &token_file).await?;
+                Ok((access_key_id, secret_access_key, Some(session_token)))
+            }
+            (_, _) => anyhow::bail!("Clap must pass valid auth parameters"),
+        }
+    }
+
+    #[cfg(unix)]
+    pub(super) async fn process_snapshot_to_s3(
+        &self,
+        progress: Progress,
+        opts: meilisearch_types::milli::update::S3SnapshotOptions,
+        mut tasks: Vec<Task>,
+    ) -> Result<Vec<Task>> {
+        use meilisearch_types::milli::update::S3SnapshotOptions;
+        use std::ffi::OsStr;
+
+        let S3SnapshotOptions {
+            s3_bucket_url,
+            s3_bucket_region,
+            s3_bucket_name,
+            s3_snapshot_prefix,
+            s3_access_key,
+            s3_secret_key,
+            s3_role_arn,
+            s3_web_identity_token_file,
+            s3_max_in_flight_parts,
+            s3_compression_level: level,
+            s3_signature_duration,
+            s3_multipart_part_size,
+        } = opts;
+
+        let must_stop_processing = self.scheduler.must_stop_processing.clone();
+        let retry_backoff = backoff::ExponentialBackoff::default();
+        let db_name = {
+            let mut base_path = self.env.path().to_owned();
+            base_path.pop();
+            base_path.file_name().and_then(OsStr::to_str).unwrap_or("data.ms").to_string()
+        };
+
+        let (reader, writer) = std::io::pipe()?;
+        let uploader_task = tokio::spawn(async move {
+            let (s3_access_key, s3_secret_key, s3_token) = Self::extract_credentials_from_options(
+                s3_access_key,
+                s3_secret_key,
+                s3_role_arn,
+                s3_web_identity_token_file,
+            )
+            .await?;
+
+            multipart_stream_to_s3(
+                s3_bucket_url,
+                s3_bucket_region,
+                s3_bucket_name,
+                s3_snapshot_prefix,
+                s3_access_key,
+                s3_secret_key,
+                s3_token,
+                s3_max_in_flight_parts,
+                s3_signature_duration,
+                s3_multipart_part_size,
+                must_stop_processing,
+                retry_backoff,
+                db_name,
+                reader,
+            )
+            .await
+        });
+
+        let index_scheduler = IndexScheduler::private_clone(self);
+        let builder_task = tokio::task::spawn_blocking(move || {
+            stream_tarball_into_pipe(progress, level, writer, index_scheduler)
+        });
+
+        let (uploader_result, builder_result) = tokio::join!(uploader_task, builder_task);
+
+        // Check uploader result first to early return on task abortion.
+        // safety: JoinHandle can return an error if the task was aborted, cancelled, or panicked.
+        uploader_result.unwrap()?;
+        builder_result.unwrap()?;
+
+        for task in &mut tasks {
+            task.status = Status::Succeeded;
+        }
+
+        Ok(tasks)
+    }
+}
+
+#[cfg(unix)]
+#[derive(Debug, Clone, serde::Deserialize)]
+struct StsCredentials {
+    #[serde(rename = "AccessKeyId")]
+    access_key_id: String,
+    #[serde(rename = "SecretAccessKey")]
+    secret_access_key: String,
+    #[serde(rename = "SessionToken")]
+    session_token: String,
+}
+
+#[cfg(unix)]
+#[derive(Debug, serde::Deserialize)]
+struct AssumeRoleWithWebIdentityResult {
+    #[serde(rename = "Credentials")]
+    credentials: StsCredentials,
+}
+
+#[cfg(unix)]
+#[derive(Debug, serde::Deserialize)]
+struct AssumeRoleWithWebIdentityResponse {
+    #[serde(rename = "AssumeRoleWithWebIdentityResult")]
+    result: AssumeRoleWithWebIdentityResult,
+}
+
+#[cfg(unix)]
+#[derive(Debug, serde::Deserialize)]
+struct StsResponse {
+    #[serde(rename = "AssumeRoleWithWebIdentityResponse")]
+    response: AssumeRoleWithWebIdentityResponse,
+}
+
+/// Streams a tarball of the database content into a pipe.
+#[cfg(unix)]
+fn stream_tarball_into_pipe(
+    progress: Progress,
+    level: u32,
+    writer: std::io::PipeWriter,
+    index_scheduler: IndexScheduler,
+) -> std::result::Result<(), Error> {
+    use std::io::Write as _;
+    use std::path::Path;
+    use std::sync::atomic::Ordering;
+
+    use meilisearch_types::milli::progress::VariableNameStep;
+    use meilisearch_types::VERSION_FILE_NAME;
+
+    use crate::processing::{AtomicUpdateFileStep, SnapshotCreationProgress};
+    use crate::scheduler::process_snapshot_creation::UPDATE_FILES_DIR_NAME;
+
+    let writer = flate2::write::GzEncoder::new(writer, flate2::Compression::new(level));
+    let mut tarball = tar::Builder::new(writer);
+
+    // 1. Snapshot the version file
+    tarball
+        .append_path_with_name(&index_scheduler.scheduler.version_file_path, VERSION_FILE_NAME)?;
+
+    // 2. Snapshot the index scheduler LMDB env
+    progress.update_progress(SnapshotCreationProgress::SnapshotTheIndexScheduler);
+    let tasks_env_file = index_scheduler.env.try_clone_inner_file()?;
+    let path = Path::new("tasks").join("data.mdb");
+    append_file_to_tarball(&mut tarball, path, tasks_env_file)?;
+
+    // 2.3 Create a read transaction on the index-scheduler
+    let rtxn = index_scheduler.env.read_txn()?;
+
+    // 2.4 Create the update files directory
+    //     And only copy the update files of the enqueued tasks
+    progress.update_progress(SnapshotCreationProgress::SnapshotTheUpdateFiles);
+    let enqueued = index_scheduler.queue.tasks.get_status(&rtxn, Status::Enqueued)?;
+    let (atomic, update_file_progress) = AtomicUpdateFileStep::new(enqueued.len() as u32);
+    progress.update_progress(update_file_progress);
+
+    // We create the update_files directory so that it
+    // always exists even if there are no update files
+    let update_files_dir = Path::new(UPDATE_FILES_DIR_NAME);
+    let src_update_files_dir = {
+        let mut path = index_scheduler.env.path().to_path_buf();
+        path.pop();
+        path.join(UPDATE_FILES_DIR_NAME)
+    };
+    tarball.append_dir(update_files_dir, src_update_files_dir)?;
+
+    for task_id in enqueued {
+        let task = index_scheduler
+            .queue
+            .tasks
+            .get_task(&rtxn, task_id)?
+            .ok_or(Error::CorruptedTaskQueue)?;
+        if let Some(content_uuid) = task.content_uuid() {
+            use std::fs::File;
+
+            let src = index_scheduler.queue.file_store.update_path(content_uuid);
+            let mut update_file = File::open(src)?;
+            let path = update_files_dir.join(content_uuid.to_string());
+            tarball.append_file(path, &mut update_file)?;
+        }
+        atomic.fetch_add(1, Ordering::Relaxed);
+    }
+
+    // 3. Snapshot every indexes
+    progress.update_progress(SnapshotCreationProgress::SnapshotTheIndexes);
+    let index_mapping = index_scheduler.index_mapper.index_mapping;
+    let nb_indexes = index_mapping.len(&rtxn)? as u32;
+    let indexes_dir = Path::new("indexes");
+    let indexes_references: Vec<_> = index_scheduler
+        .index_mapper
+        .index_mapping
+        .iter(&rtxn)?
+        .map(|res| res.map_err(Error::from).map(|(name, uuid)| (name.to_string(), uuid)))
+        .collect::<Result<_, Error>>()?;
+
+    // It's prettier to use a for loop instead of the IndexMapper::try_for_each_index
+    // method, especially when we need to access the UUID, local path and index number.
+    for (i, (name, uuid)) in indexes_references.into_iter().enumerate() {
+        progress.update_progress(VariableNameStep::<SnapshotCreationProgress>::new(
+            &name, i as u32, nb_indexes,
+        ));
+        let path = indexes_dir.join(uuid.to_string()).join("data.mdb");
+        let index = index_scheduler.index_mapper.index(&rtxn, &name)?;
+        let index_file = index.try_clone_inner_file()?;
+        tracing::trace!("Appending index file for {name} in {}", path.display());
+        append_file_to_tarball(&mut tarball, path, index_file)?;
+    }
+
+    drop(rtxn);
+
+    // 4. Snapshot the auth LMDB env
+    progress.update_progress(SnapshotCreationProgress::SnapshotTheApiKeys);
+    let auth_env_file = index_scheduler.scheduler.auth_env.try_clone_inner_file()?;
+    let path = Path::new("auth").join("data.mdb");
+    append_file_to_tarball(&mut tarball, path, auth_env_file)?;
+
+    let mut gzencoder = tarball.into_inner()?;
+    gzencoder.flush()?;
+    gzencoder.try_finish()?;
+    let mut writer = gzencoder.finish()?;
+    writer.flush()?;
+
+    Result::<_, Error>::Ok(())
+}
+
+#[cfg(unix)]
+fn append_file_to_tarball<W, P>(
+    tarball: &mut tar::Builder<W>,
+    path: P,
+    mut auth_env_file: std::fs::File,
+) -> Result<(), Error>
+where
+    W: std::io::Write,
+    P: AsRef<std::path::Path>,
+{
+    use std::io::{Seek as _, SeekFrom};
+
+    // Note: A previous snapshot operation may have left the cursor
+    //       at the end of the file so we need to seek to the start.
+    auth_env_file.seek(SeekFrom::Start(0))?;
+    tarball.append_file(path, &mut auth_env_file)?;
+    Ok(())
+}
+
+/// Streams the content read from the given reader to S3.
+#[allow(clippy::too_many_arguments)]
+#[cfg(unix)]
+async fn multipart_stream_to_s3(
+    s3_bucket_url: String,
+    s3_bucket_region: String,
+    s3_bucket_name: String,
+    s3_snapshot_prefix: String,
+    s3_access_key: String,
+    s3_secret_key: String,
+    s3_token: Option<String>,
+    s3_max_in_flight_parts: std::num::NonZero<usize>,
+    s3_signature_duration: std::time::Duration,
+    s3_multipart_part_size: u64,
+    must_stop_processing: super::MustStopProcessing,
+    retry_backoff: backoff::exponential::ExponentialBackoff<backoff::SystemClock>,
+    db_name: String,
+    reader: std::io::PipeReader,
+) -> Result<(), Error> {
+    use std::collections::VecDeque;
+    use std::io;
+    use std::os::fd::OwnedFd;
+    use std::path::PathBuf;
+
+    use bytes::{Bytes, BytesMut};
+    use reqwest::{Client, Response};
+    use rusty_s3::actions::CreateMultipartUpload;
+    use rusty_s3::{Bucket, BucketError, Credentials, S3Action as _, UrlStyle};
+    use tokio::task::JoinHandle;
+
+    let reader = OwnedFd::from(reader);
+    let reader = tokio::net::unix::pipe::Receiver::from_owned_fd(reader)?;
+    let s3_snapshot_prefix = PathBuf::from(s3_snapshot_prefix);
+    let url =
+        s3_bucket_url.parse().map_err(BucketError::ParseError).map_err(Error::S3BucketError)?;
+    let bucket = Bucket::new(url, UrlStyle::Path, s3_bucket_name, s3_bucket_region)
+        .map_err(Error::S3BucketError)?;
+    let credential = match s3_token {
+        Some(token) => Credentials::new_with_token(s3_access_key, s3_secret_key, token),
+        None => Credentials::new(s3_access_key, s3_secret_key),
+    };
+
+    // Note for the future (rust 1.91+): use with_added_extension, it's prettier
+    let object_path = s3_snapshot_prefix.join(format!("{db_name}.snapshot"));
+    // Note: It doesn't work on Windows and if a port to this platform is needed,
+    //       use the slash-path crate or similar to get the correct path separator.
+    let object = object_path.display().to_string();
+
+    let action = bucket.create_multipart_upload(Some(&credential), &object);
+    let url = action.sign(s3_signature_duration);
+
+    let client = Client::new();
+    let resp = client.post(url).send().await.map_err(Error::S3HttpError)?;
+    let status = resp.status();
+
+    let body = match resp.error_for_status_ref() {
+        Ok(_) => resp.text().await.map_err(Error::S3HttpError)?,
+        Err(_) => {
+            return Err(Error::S3Error { status, body: resp.text().await.unwrap_or_default() })
+        }
+    };
+
+    let multipart =
+        CreateMultipartUpload::parse_response(&body).map_err(|e| Error::S3XmlError(Box::new(e)))?;
+    tracing::debug!("Starting the upload of the snapshot to {object}");
+
+    // We use this bumpalo for etags strings.
+    let bump = bumpalo::Bump::new();
+    let mut etags = Vec::<&str>::new();
+    let mut in_flight = VecDeque::<(JoinHandle<reqwest::Result<Response>>, Bytes)>::with_capacity(
+        s3_max_in_flight_parts.get(),
+    );
+
+    // Part numbers start at 1 and cannot be larger than 10k
+    for part_number in 1u16.. {
+        if must_stop_processing.get() {
+            return Err(Error::AbortedTask);
+        }
+
+        let part_upload =
+            bucket.upload_part(Some(&credential), &object, part_number, multipart.upload_id());
+        let url = part_upload.sign(s3_signature_duration);
+
+        // Wait for a buffer to be ready if there are in-flight parts that landed
+        let mut buffer = if in_flight.len() >= s3_max_in_flight_parts.get() {
+            let (handle, buffer) = in_flight.pop_front().expect("At least one in flight request");
+            let resp = join_and_map_error(handle).await?;
+            extract_and_append_etag(&bump, &mut etags, resp.headers())?;
+
+            let mut buffer = match buffer.try_into_mut() {
+                Ok(buffer) => buffer,
+                Err(_) => unreachable!("All bytes references were consumed in the task"),
+            };
+            buffer.clear();
+            buffer
+        } else {
+            BytesMut::with_capacity(s3_multipart_part_size as usize)
+        };
+
+        // If we successfully read enough bytes,
+        // we can continue and send the buffer/part
+        while buffer.len() < (s3_multipart_part_size as usize / 2) {
+            // Wait for the pipe to be readable
+
+            reader.readable().await?;
+
+            match reader.try_read_buf(&mut buffer) {
+                Ok(0) => break,
+                // We read some bytes but maybe not enough
+                Ok(_) => continue,
+                // The readiness event is a false positive.
+                Err(ref e) if e.kind() == io::ErrorKind::WouldBlock => continue,
+                Err(e) => return Err(e.into()),
+            }
+        }
+
+        if buffer.is_empty() {
+            // Break the loop if the buffer is
+            // empty after we tried to read bytes
+            break;
+        }
+
+        let body = buffer.freeze();
+        tracing::trace!("Sending part {part_number}");
+        let task = tokio::spawn({
+            let client = client.clone();
+            let body = body.clone();
+            backoff::future::retry(retry_backoff.clone(), move || {
+                let client = client.clone();
+                let url = url.clone();
+                let body = body.clone();
+                async move {
+                    match client.put(url).body(body).send().await {
+                        Ok(resp) if resp.status().is_client_error() => {
+                            resp.error_for_status().map_err(backoff::Error::Permanent)
+                        }
+                        Ok(resp) => Ok(resp),
+                        Err(e) => Err(backoff::Error::transient(e)),
+                    }
+                }
+            })
+        });
+        in_flight.push_back((task, body));
+    }
+
+    for (handle, _buffer) in in_flight {
+        let resp = join_and_map_error(handle).await?;
+        extract_and_append_etag(&bump, &mut etags, resp.headers())?;
+    }
+
+    tracing::debug!("Finalizing the multipart upload");
+
+    let action = bucket.complete_multipart_upload(
+        Some(&credential),
+        &object,
+        multipart.upload_id(),
+        etags.iter().map(AsRef::as_ref),
+    );
+    let url = action.sign(s3_signature_duration);
+    let body = action.body();
+    let resp = backoff::future::retry(retry_backoff, move || {
+        let client = client.clone();
+        let url = url.clone();
+        let body = body.clone();
+        async move {
+            match client.post(url).body(body).send().await {
+                Ok(resp) if resp.status().is_client_error() => {
+                    Err(backoff::Error::Permanent(Error::S3Error {
+                        status: resp.status(),
+                        body: resp.text().await.unwrap_or_default(),
+                    }))
+                }
+                Ok(resp) => Ok(resp),
+                Err(e) => Err(backoff::Error::transient(Error::S3HttpError(e))),
+            }
+        }
+    })
+    .await?;
+
+    let status = resp.status();
+    let body = resp.text().await.map_err(|e| Error::S3Error { status, body: e.to_string() })?;
+    if status.is_success() {
+        Ok(())
+    } else {
+        Err(Error::S3Error { status, body })
+    }
+}
+
+#[cfg(unix)]
+async fn join_and_map_error(
+    join_handle: tokio::task::JoinHandle<Result<reqwest::Response, reqwest::Error>>,
+) -> Result<reqwest::Response> {
+    // safety: Panic happens if the task (JoinHandle) was aborted, cancelled, or panicked
+    let request = join_handle.await.unwrap();
+    let resp = request.map_err(Error::S3HttpError)?;
+    match resp.error_for_status_ref() {
+        Ok(_) => Ok(resp),
+        Err(_) => Err(Error::S3Error {
+            status: resp.status(),
+            body: resp.text().await.unwrap_or_default(),
+        }),
+    }
+}
+
+#[cfg(unix)]
+fn extract_and_append_etag<'b>(
+    bump: &'b bumpalo::Bump,
+    etags: &mut Vec<&'b str>,
+    headers: &reqwest::header::HeaderMap,
+) -> Result<()> {
+    use reqwest::header::ETAG;
+
+    let etag = headers.get(ETAG).ok_or_else(|| Error::S3XmlError("Missing ETag header".into()))?;
+    let etag = etag.to_str().map_err(|e| Error::S3XmlError(Box::new(e)))?;
+    etags.push(bump.alloc_str(etag));
+
+    Ok(())
+}

crates/index-scheduler/src/scheduler/mod.rs

@@ -1,7 +1,12 @@
 mod autobatcher;
 #[cfg(test)]
 mod autobatcher_test;
+#[cfg(not(feature = "enterprise"))]
+mod community_edition;
 mod create_batch;
+#[cfg(feature = "enterprise")]
+mod enterprise_edition;
+
 mod process_batch;
 mod process_dump_creation;
 mod process_export;
@@ -21,7 +26,6 @@ use std::path::PathBuf;
 use std::sync::atomic::{AtomicBool, AtomicU32, Ordering};
 use std::sync::Arc;
 
-use convert_case::{Case, Casing as _};
 use meilisearch_types::error::ResponseError;
 use meilisearch_types::heed::{Env, WithoutTls};
 use meilisearch_types::milli;
@@ -133,6 +137,7 @@ impl Scheduler {
             max_number_of_tasks: _,
             max_number_of_batched_tasks,
             batched_tasks_size_limit,
+            export_default_payload_size_bytes: _,
             instance_features: _,
             auto_upgrade: _,
             embedding_cache_cap,
@@ -178,6 +183,8 @@ impl IndexScheduler {
             self.breakpoint(crate::test_utils::Breakpoint::Start);
         }
 
+        let previous_processing_batch = self.processing_tasks.write().unwrap().stop_processing();
+
         if self.cleanup_enabled {
             let mut wtxn = self.env.write_txn()?;
             self.queue.cleanup_task_queue(&mut wtxn)?;
@@ -185,10 +192,15 @@ impl IndexScheduler {
         }
 
         let rtxn = self.env.read_txn().map_err(Error::HeedTransaction)?;
-        let (batch, mut processing_batch) =
+        let (batch, mut processing_batch) = match self
-            match self.create_next_batch(&rtxn).map_err(|e| Error::CreateBatch(Box::new(e)))? {
+            .create_next_batch(&rtxn, &previous_processing_batch.processing)
+            .map_err(|e| Error::CreateBatch(Box::new(e)))?
+        {
             Some(batch) => batch,
-                None => return Ok(TickOutcome::WaitForSignal),
+            None => {
+                *self.processing_tasks.write().unwrap() = previous_processing_batch;
+                return Ok(TickOutcome::WaitForSignal);
+            }
         };
         let index_uid = batch.index_uid().map(ToOwned::to_owned);
         drop(rtxn);
@@ -219,7 +231,12 @@ impl IndexScheduler {
                 let handle = std::thread::Builder::new()
                     .name(String::from("batch-operation"))
                     .spawn_scoped(s, move || {
-                        cloned_index_scheduler.process_batch(batch, processing_batch, p)
+                        cloned_index_scheduler.process_batch(
+                            batch,
+                            processing_batch,
+                            p,
+                            &self.network(),
+                        )
                     })
                     .unwrap();
 
@@ -260,7 +277,14 @@ impl IndexScheduler {
         self.maybe_fail(crate::test_utils::FailureLocation::AcquiringWtxn)?;
 
         progress.update_progress(BatchProgress::WritingTasksToDisk);
+
         processing_batch.finished();
+        // whether the batch made progress.
+        // a batch make progress if it failed or if it contains at least one fully processed (or cancelled) task.
+        //
+        // if a batch did not make progress, it means that all of its tasks are waiting on the scheduler to make progress,
+        // and so we must wait for new tasks. Such a batch is not persisted to DB, and is resumed on the next tick.
+        let mut batch_made_progress = false;
         let mut stop_scheduler_forever = false;
         let mut wtxn = self.env.write_txn().map_err(Error::HeedTransaction)?;
         let mut canceled = RoaringBitmap::new();
@@ -281,7 +305,11 @@ impl IndexScheduler {
                 #[allow(unused_variables)]
                 for (i, mut task) in tasks.into_iter().enumerate() {
                     task_progress.fetch_add(1, Ordering::Relaxed);
-                    processing_batch.update(&mut task);
+                    processing_batch.update_from_task(&task);
+                    if !matches!(task.status, Status::Processing | Status::Enqueued) {
+                        batch_made_progress = true;
+                        processing_batch.finish_task(&mut task);
+                    }
                     if task.status == Status::Canceled {
                         canceled.insert(task.uid);
                         canceled_by = task.canceled_by;
@@ -348,6 +376,9 @@ impl IndexScheduler {
             }
             // In case of a failure we must get back and patch all the tasks with the error.
             Err(err) => {
+                // always persist failed batches
+                batch_made_progress = true;
+
                 #[cfg(test)]
                 self.breakpoint(crate::test_utils::Breakpoint::ProcessBatchFailed);
                 let (task_progress, task_progress_obj) = AtomicTaskStep::new(ids.len() as u32);
@@ -371,7 +402,10 @@ impl IndexScheduler {
                     task.status = Status::Failed;
                     task.error = Some(error.clone());
                     task.details = task.details.map(|d| d.to_failed());
-                    processing_batch.update(&mut task);
+                    processing_batch.update_from_task(&task);
+                    if !matches!(task.status, Status::Processing | Status::Enqueued) {
+                        processing_batch.finish_task(&mut task);
+                    }
 
                     #[cfg(test)]
                     self.maybe_fail(
@@ -394,44 +428,12 @@ impl IndexScheduler {
         let ProcessBatchInfo { congestion, pre_commit_dabases_sizes, post_commit_dabases_sizes } =
             process_batch_info;
 
-        processing_batch.stats.progress_trace =
+        processing_batch.write_stats(
-            progress.accumulated_durations().into_iter().map(|(k, v)| (k, v.into())).collect();
+            &progress,
-        processing_batch.stats.write_channel_congestion = congestion.map(|congestion| {
+            congestion,
-            let mut congestion_info = serde_json::Map::new();
+            pre_commit_dabases_sizes,
-            congestion_info.insert("attempts".into(), congestion.attempts.into());
+            post_commit_dabases_sizes,
-            congestion_info.insert("blocking_attempts".into(), congestion.blocking_attempts.into());
+        );
-            congestion_info.insert("blocking_ratio".into(), congestion.congestion_ratio().into());
-            congestion_info
-        });
-        processing_batch.stats.internal_database_sizes = pre_commit_dabases_sizes
-            .iter()
-            .flat_map(|(dbname, pre_size)| {
-                post_commit_dabases_sizes
-                    .get(dbname)
-                    .map(|post_size| {
-                        use std::cmp::Ordering::{Equal, Greater, Less};
-
-                        use byte_unit::Byte;
-                        use byte_unit::UnitType::Binary;
-
-                        let post = Byte::from_u64(*post_size as u64).get_appropriate_unit(Binary);
-                        let diff_size = post_size.abs_diff(*pre_size) as u64;
-                        let diff = Byte::from_u64(diff_size).get_appropriate_unit(Binary);
-                        let sign = match post_size.cmp(pre_size) {
-                            Equal => return None,
-                            Greater => "+",
-                            Less => "-",
-                        };
-
-                        Some((
-                            dbname.to_case(Case::Camel),
-                            format!("{post:#.2} ({sign}{diff:#.2})").into(),
-                        ))
-                    })
-                    .into_iter()
-                    .flatten()
-            })
-            .collect();
 
         if let Some(congestion) = congestion {
             tracing::debug!(
@@ -444,13 +446,16 @@ impl IndexScheduler {
 
         tracing::debug!("call trace: {:?}", progress.accumulated_durations());
 
+        if batch_made_progress {
             self.queue.write_batch(&mut wtxn, processing_batch, &ids)?;
+        }
 
         #[cfg(test)]
         self.maybe_fail(crate::test_utils::FailureLocation::CommittingWtxn)?;
 
         wtxn.commit().map_err(Error::HeedTransaction)?;
 
+        if batch_made_progress {
             // We should stop processing AFTER everything is processed and written to disk otherwise, a batch (which only lives in RAM) may appear in the processing task
             // and then become « not found » for some time until the commit everything is written and the final commit is made.
             self.processing_tasks.write().unwrap().stop_processing();
@@ -480,10 +485,10 @@ impl IndexScheduler {
             })?;
 
             self.notify_webhooks(ids);
+        }
 
         #[cfg(test)]
         self.breakpoint(crate::test_utils::Breakpoint::AfterProcessing);
-
         if stop_scheduler_forever {
             Ok(TickOutcome::StopProcessingForever)
         } else {

crates/index-scheduler/src/scheduler/process_batch.rs

@@ -10,6 +10,7 @@ use meilisearch_types::heed::{RoTxn, RwTxn};
 use meilisearch_types::milli::heed::CompactionOption;
 use meilisearch_types::milli::progress::{Progress, VariableNameStep};
 use meilisearch_types::milli::{self, ChannelCongestion};
+use meilisearch_types::network::Network;
 use meilisearch_types::tasks::{Details, IndexSwap, Kind, KindWithContent, Status, Task};
 use meilisearch_types::versioning::{VERSION_MAJOR, VERSION_MINOR, VERSION_PATCH};
 use milli::update::Settings as MilliSettings;
@@ -55,6 +56,7 @@ impl IndexScheduler {
         batch: Batch,
         current_batch: &mut ProcessingBatch,
         progress: Progress,
+        network: &Network,
     ) -> Result<(Vec<Task>, ProcessBatchInfo)> {
         #[cfg(test)]
         {
@@ -176,6 +178,7 @@ impl IndexScheduler {
                     op,
                     &progress,
                     current_batch.embedder_stats.clone(),
+                    network,
                 )?;
 
                 {
@@ -235,6 +238,7 @@ impl IndexScheduler {
                     Batch::IndexUpdate { index_uid, primary_key, new_index_uid: None, task },
                     current_batch,
                     progress,
+                    network,
                 )
             }
             Batch::IndexUpdate { index_uid, primary_key, new_index_uid, mut task } => {
@@ -539,6 +543,10 @@ impl IndexScheduler {
 
                 Ok((tasks, ProcessBatchInfo::default()))
             }
+            Batch::NetworkIndexBatch { network_task, inner_batch } => {
+                self.process_network_index_batch(network_task, inner_batch, current_batch, progress)
+            }
+            Batch::NetworkReady { task } => self.process_network_ready(task, progress),
         }
     }
 

crates/index-scheduler/src/scheduler/process_export.rs

@@ -1,5 +1,6 @@
 use std::collections::BTreeMap;
 use std::io::{self, Write as _};
+use std::ops::ControlFlow;
 use std::sync::atomic;
 use std::time::Duration;
 
@@ -7,6 +8,7 @@ use backoff::ExponentialBackoff;
 use byte_unit::Byte;
 use flate2::write::GzEncoder;
 use flate2::Compression;
+use meilisearch_types::error::Code;
 use meilisearch_types::index_uid_pattern::IndexUidPattern;
 use meilisearch_types::milli::constants::RESERVED_VECTORS_FIELD_NAME;
 use meilisearch_types::milli::index::EmbeddingsWithMetadata;
@@ -15,7 +17,10 @@ use meilisearch_types::milli::update::{request_threads, Setting};
 use meilisearch_types::milli::vector::parsed_vectors::{ExplicitVectors, VectorOrArrayOfVectors};
 use meilisearch_types::milli::{self, obkv_to_json, Filter, InternalError};
 use meilisearch_types::settings::{self, SecretPolicy};
+use meilisearch_types::tasks::network::headers::SetHeader as _;
+use meilisearch_types::tasks::network::{headers, ImportData, ImportMetadata, Origin};
 use meilisearch_types::tasks::{DetailsExportIndexSettings, ExportIndexSettings};
+use roaring::RoaringBitmap;
 use serde::Deserialize;
 use ureq::{json, Response};
 
@@ -50,6 +55,7 @@ impl IndexScheduler {
         let agent = ureq::AgentBuilder::new().timeout(Duration::from_secs(5)).build();
         let must_stop_processing = self.scheduler.must_stop_processing.clone();
         for (i, (_pattern, uid, export_settings)) in indexes.iter().enumerate() {
+            let err = |err| Error::from_milli(err, Some(uid.to_string()));
             if must_stop_processing.get() {
                 return Err(Error::AbortedTask);
             }
@@ -61,14 +67,62 @@ impl IndexScheduler {
             ));
 
             let ExportIndexSettings { filter, override_settings } = export_settings;
+
             let index = self.index(uid)?;
             let index_rtxn = index.read_txn()?;
-            let bearer = api_key.map(|api_key| format!("Bearer {api_key}"));
+            let filter = filter.as_ref().map(Filter::from_json).transpose().map_err(err)?.flatten();
+            let filter_universe =
+                filter.map(|f| f.evaluate(&index_rtxn, &index)).transpose().map_err(err)?;
+            let whole_universe =
+                index.documents_ids(&index_rtxn).map_err(milli::Error::from).map_err(err)?;
+            let universe = filter_universe.unwrap_or(whole_universe);
+            let target = TargetInstance { remote_name: None, base_url, api_key };
+            let ctx = ExportContext {
+                index: &index,
+                index_rtxn: &index_rtxn,
+                universe: &universe,
+                progress: &progress,
+                agent: &agent,
+                must_stop_processing: &must_stop_processing,
+            };
+            let options = ExportOptions {
+                index_uid: uid,
+                payload_size,
+                override_settings: *override_settings,
+                export_mode: ExportMode::ExportRoute,
+            };
+            let total_documents = self.export_one_index(target, options, ctx)?;
 
-            // First, check if the index already exists
+            output.insert(
-            let url = format!("{base_url}/indexes/{uid}");
+                IndexUidPattern::new_unchecked(uid.clone()),
-            let response = retry(&must_stop_processing, || {
+                DetailsExportIndexSettings {
-                let mut request = agent.get(&url);
+                    settings: (*export_settings).clone(),
+                    matched_documents: Some(total_documents),
+                },
+            );
+        }
+
+        Ok(output)
+    }
+
+    pub(super) fn export_one_index(
+        &self,
+        target: TargetInstance<'_>,
+        options: ExportOptions<'_>,
+        ctx: ExportContext<'_>,
+    ) -> Result<u64, Error> {
+        let err = |err| Error::from_milli(err, Some(options.index_uid.to_string()));
+        let total_index_documents = ctx.universe.len();
+        let task_network = options.task_network(total_index_documents);
+
+        let bearer = target.api_key.map(|api_key| format!("Bearer {api_key}"));
+        let url = format!(
+            "{base_url}/indexes/{index_uid}",
+            base_url = target.base_url,
+            index_uid = options.index_uid
+        );
+        let response = retry(ctx.must_stop_processing, || {
+            let mut request = ctx.agent.get(&url);
             if let Some(bearer) = &bearer {
                 request = request.set("Authorization", bearer);
             }
@@ -77,126 +131,146 @@ impl IndexScheduler {
         });
         let index_exists = match response {
             Ok(response) => response.status() == 200,
-                Err(Error::FromRemoteWhenExporting { code, .. }) if code == "index_not_found" => {
+            Err(Error::FromRemoteWhenExporting { code, .. })
+                if code == Code::IndexNotFound.name() =>
+            {
                 false
             }
             Err(e) => return Err(e),
         };
-
+        let primary_key =
-            let primary_key = index
+            ctx.index.primary_key(ctx.index_rtxn).map_err(milli::Error::from).map_err(err)?;
-                .primary_key(&index_rtxn)
-                .map_err(|e| Error::from_milli(e.into(), Some(uid.to_string())))?;
-
-            // Create the index
         if !index_exists {
-                let url = format!("{base_url}/indexes");
+            let url = format!("{base_url}/indexes", base_url = target.base_url);
-                retry(&must_stop_processing, || {
+            let _ = handle_response(
-                    let mut request = agent.post(&url);
+                target.remote_name,
-                    if let Some(bearer) = &bearer {
+                retry(ctx.must_stop_processing, || {
+                    let mut request = ctx.agent.post(&url);
+
+                    if let Some((import_data, origin, metadata)) = &task_network {
+                        request = set_network_ureq_headers(request, import_data, origin, metadata);
+                    }
+
+                    if let Some(bearer) = bearer.as_ref() {
                         request = request.set("Authorization", bearer);
                     }
-                    let index_param = json!({ "uid": uid, "primaryKey": primary_key });
+                    let index_param =
-                    request.send_json(&index_param).map_err(into_backoff_error)
+                        json!({ "uid": options.index_uid, "primaryKey": primary_key });
-                })?;
-            }
 
-            // Patch the index primary key
+                    request.send_json(&index_param).map_err(into_backoff_error)
-            if index_exists && *override_settings {
+                }),
-                let url = format!("{base_url}/indexes/{uid}");
+            )?;
-                retry(&must_stop_processing, || {
+        }
-                    let mut request = agent.patch(&url);
+        if index_exists && options.override_settings {
+            let _ = handle_response(
+                target.remote_name,
+                retry(ctx.must_stop_processing, || {
+                    let mut request = ctx.agent.patch(&url);
+                    if let Some((import_data, origin, metadata)) = &task_network {
+                        request = set_network_ureq_headers(request, import_data, origin, metadata);
+                    }
                     if let Some(bearer) = &bearer {
                         request = request.set("Authorization", bearer);
                     }
                     let index_param = json!({ "primaryKey": primary_key });
                     request.send_json(&index_param).map_err(into_backoff_error)
-                })?;
+                }),
+            )?;
         }
-
+        if !index_exists || options.override_settings {
-            // Send the index settings
-            if !index_exists || *override_settings {
             let mut settings =
-                    settings::settings(&index, &index_rtxn, SecretPolicy::RevealSecrets)
+                settings::settings(ctx.index, ctx.index_rtxn, SecretPolicy::RevealSecrets)
-                        .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?;
+                    .map_err(err)?;
             // Remove the experimental chat setting if not enabled
             if self.features().check_chat_completions("exporting chat settings").is_err() {
                 settings.chat = Setting::NotSet;
             }
             // Retry logic for sending settings
-                let url = format!("{base_url}/indexes/{uid}/settings");
+            let url = format!(
-                retry(&must_stop_processing, || {
+                "{base_url}/indexes/{index_uid}/settings",
-                    let mut request = agent.patch(&url);
+                base_url = target.base_url,
+                index_uid = options.index_uid
+            );
+
+            let _ = handle_response(
+                target.remote_name,
+                retry(ctx.must_stop_processing, || {
+                    let mut request = ctx.agent.patch(&url);
+
+                    if let Some((import_data, origin, metadata)) = &task_network {
+                        request = set_network_ureq_headers(request, import_data, origin, metadata);
+                    }
+
                     if let Some(bearer) = bearer.as_ref() {
                         request = request.set("Authorization", bearer);
                     }
                     request.send_json(settings.clone()).map_err(into_backoff_error)
-                })?;
+                }),
+            )?;
         }
 
-            let filter = filter
+        let fields_ids_map = ctx.index.fields_ids_map(ctx.index_rtxn)?;
-                .as_ref()
-                .map(Filter::from_json)
-                .transpose()
-                .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?
-                .flatten();
-
-            let filter_universe = filter
-                .map(|f| f.evaluate(&index_rtxn, &index))
-                .transpose()
-                .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?;
-            let whole_universe = index
-                .documents_ids(&index_rtxn)
-                .map_err(|e| Error::from_milli(e.into(), Some(uid.to_string())))?;
-            let universe = filter_universe.unwrap_or(whole_universe);
-
-            let fields_ids_map = index.fields_ids_map(&index_rtxn)?;
         let all_fields: Vec<_> = fields_ids_map.iter().map(|(id, _)| id).collect();
-
+        let total_documents = ctx.universe.len() as u32;
-            // We don't need to keep this one alive as we will
-            // spawn many threads to process the documents
-            drop(index_rtxn);
-
-            let total_documents = universe.len() as u32;
         let (step, progress_step) = AtomicDocumentStep::new(total_documents);
-            progress.update_progress(progress_step);
+        ctx.progress.update_progress(progress_step);
 
-            output.insert(
+        let limit = options
-                IndexUidPattern::new_unchecked(uid.clone()),
+            .payload_size
-                DetailsExportIndexSettings {
+            .map(|ps| ps.as_u64() as usize)
-                    settings: (*export_settings).clone(),
+            .unwrap_or(self.export_default_payload_size_bytes.as_u64() as usize);
-                    matched_documents: Some(total_documents as u64),
+        let documents_url = format!(
-                },
+            "{base_url}/indexes/{index_uid}/documents",
+            base_url = target.base_url,
+            index_uid = options.index_uid
         );
 
-            let limit = payload_size.map(|ps| ps.as_u64() as usize).unwrap_or(20 * 1024 * 1024); // defaults to 20 MiB
+        // no document to send, but we must still send a task when performing network balancing
-            let documents_url = format!("{base_url}/indexes/{uid}/documents");
+        if ctx.universe.is_empty() {
+            if let Some((import_data, network_change_origin, metadata)) = task_network {
+                let mut compressed_buffer = Vec::new();
+                // ignore control flow, we're returning anyway
+                let _ = send_buffer(
+                    b" ", // needs something otherwise meili complains about missing payload
+                    &mut compressed_buffer,
+                    ctx.must_stop_processing,
+                    ctx.agent,
+                    &documents_url,
+                    target.remote_name,
+                    bearer.as_deref(),
+                    Some(&(import_data, network_change_origin.clone(), metadata)),
+                    &err,
+                )?;
+            }
+            return Ok(0);
+        }
 
         let results = request_threads()
-                .broadcast(|ctx| {
+            .broadcast(|broadcast| {
-                    let index_rtxn = index
+                let mut task_network = options.task_network(total_index_documents);
-                        .read_txn()
+
-                        .map_err(|e| Error::from_milli(e.into(), Some(uid.to_string())))?;
+                let index_rtxn = ctx.index.read_txn().map_err(milli::Error::from).map_err(err)?;
 
                 let mut buffer = Vec::new();
                 let mut tmp_buffer = Vec::new();
                 let mut compressed_buffer = Vec::new();
-                    for (i, docid) in universe.iter().enumerate() {
+                for (i, docid) in ctx.universe.iter().enumerate() {
-                        if i % ctx.num_threads() != ctx.index() {
+                    if i % broadcast.num_threads() != broadcast.index() {
                         continue;
                     }
+                    if let Some((import_data, _, metadata)) = &mut task_network {
+                        import_data.document_count += 1;
+                        metadata.task_key = Some(docid);
+                    }
 
-                        let document = index
+                    let document = ctx.index.document(&index_rtxn, docid).map_err(err)?;
-                            .document(&index_rtxn, docid)
-                            .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?;
 
-                        let mut document = obkv_to_json(&all_fields, &fields_ids_map, document)
+                    let mut document =
-                            .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?;
+                        obkv_to_json(&all_fields, &fields_ids_map, document).map_err(err)?;
 
                     // TODO definitely factorize this code
                     'inject_vectors: {
-                            let embeddings = index
+                        let embeddings = ctx.index.embeddings(&index_rtxn, docid).map_err(err)?;
-                                .embeddings(&index_rtxn, docid)
-                                .map_err(|e| Error::from_milli(e, Some(uid.to_string())))?;
 
                         if embeddings.is_empty() {
                             break 'inject_vectors;
@@ -207,15 +281,12 @@ impl IndexScheduler {
                             .or_insert(serde_json::Value::Object(Default::default()));
 
                         let serde_json::Value::Object(vectors) = vectors else {
-                                return Err(Error::from_milli(
+                            return Err(err(milli::Error::UserError(
-                                    milli::Error::UserError(
                                 milli::UserError::InvalidVectorsMapType {
                                     document_id: {
-                                                if let Ok(Some(Ok(index))) = index
+                                        if let Ok(Some(Ok(index))) = ctx
-                                                    .external_id_of(
+                                            .index
-                                                        &index_rtxn,
+                                            .external_id_of(&index_rtxn, std::iter::once(docid))
-                                                        std::iter::once(docid),
-                                                    )
                                             .map(|it| it.into_iter().next())
                                         {
                                             index
@@ -225,9 +296,7 @@ impl IndexScheduler {
                                     },
                                     value: vectors.clone(),
                                 },
-                                    ),
+                            )));
-                                    Some(uid.to_string()),
-                                ));
                         };
 
                         for (
@@ -236,9 +305,9 @@ impl IndexScheduler {
                         ) in embeddings
                         {
                             let embeddings = ExplicitVectors {
-                                    embeddings: Some(
+                                embeddings: Some(VectorOrArrayOfVectors::from_array_of_vectors(
-                                        VectorOrArrayOfVectors::from_array_of_vectors(embeddings),
+                                    embeddings,
-                                    ),
+                                )),
                                 regenerate: regenerate &&
                                 // Meilisearch does not handle well dumps with fragments, because as the fragments
                                 // are marked as user-provided,
@@ -246,42 +315,40 @@ impl IndexScheduler {
                                 // To prevent this, we mark embeddings has non regenerate in this case.
                                 !has_fragments,
                             };
-                                vectors.insert(
+                            vectors
-                                    embedder_name,
+                                .insert(embedder_name, serde_json::to_value(embeddings).unwrap());
-                                    serde_json::to_value(embeddings).unwrap(),
-                                );
                         }
                     }
 
                     tmp_buffer.clear();
                     serde_json::to_writer(&mut tmp_buffer, &document)
                         .map_err(milli::InternalError::from)
-                            .map_err(|e| Error::from_milli(e.into(), Some(uid.to_string())))?;
+                        .map_err(milli::Error::from)
+                        .map_err(err)?;
 
                     // Make sure we put at least one document in the buffer even
                     // though we might go above the buffer limit before sending
                     if !buffer.is_empty() && buffer.len() + tmp_buffer.len() > limit {
-                            // We compress the documents before sending them
+                        let control_flow = send_buffer(
-                            let mut encoder =
+                            &buffer,
-                                GzEncoder::new(&mut compressed_buffer, Compression::default());
+                            &mut compressed_buffer,
-                            encoder
+                            ctx.must_stop_processing,
-                                .write_all(&buffer)
+                            ctx.agent,
-                                .map_err(|e| Error::from_milli(e.into(), Some(uid.clone())))?;
+                            &documents_url,
-                            encoder
+                            target.remote_name,
-                                .finish()
+                            bearer.as_deref(),
-                                .map_err(|e| Error::from_milli(e.into(), Some(uid.clone())))?;
+                            task_network.as_ref(),
-
+                            &err,
-                            retry(&must_stop_processing, || {
+                        )?;
-                                let mut request = agent.post(&documents_url);
-                                request = request.set("Content-Type", "application/x-ndjson");
-                                request = request.set("Content-Encoding", "gzip");
-                                if let Some(bearer) = &bearer {
-                                    request = request.set("Authorization", bearer);
-                                }
-                                request.send_bytes(&compressed_buffer).map_err(into_backoff_error)
-                            })?;
                         buffer.clear();
                         compressed_buffer.clear();
+                        if let Some((import_data, _, metadata)) = &mut task_network {
+                            import_data.document_count = 0;
+                            metadata.task_key = None;
+                        }
+                        if control_flow.is_break() {
+                            return Ok(());
+                        }
                     }
                     buffer.extend_from_slice(&tmp_buffer);
 
@@ -290,31 +357,183 @@ impl IndexScheduler {
                     }
                 }
 
-                    retry(&must_stop_processing, || {
+                // send the last buffered documents if any
-                        let mut request = agent.post(&documents_url);
+                if !buffer.is_empty() {
-                        request = request.set("Content-Type", "application/x-ndjson");
+                    // ignore control flow here
-                        if let Some(bearer) = &bearer {
+                    let _ = send_buffer(
-                            request = request.set("Authorization", bearer);
+                        &buffer,
+                        &mut compressed_buffer,
+                        ctx.must_stop_processing,
+                        ctx.agent,
+                        &documents_url,
+                        target.remote_name,
+                        bearer.as_deref(),
+                        task_network.as_ref(),
+                        &err,
+                    )?;
                 }
-                        request.send_bytes(&buffer).map_err(into_backoff_error)
-                    })?;
 
                 Ok(())
             })
-                .map_err(|e| {
+            .map_err(|e| err(milli::Error::InternalError(InternalError::PanicInThreadPool(e))))?;
-                    Error::from_milli(
-                        milli::Error::InternalError(InternalError::PanicInThreadPool(e)),
-                        Some(uid.to_string()),
-                    )
-                })?;
         for result in results {
             result?;
         }
-
         step.store(total_documents, atomic::Ordering::Relaxed);
+        Ok(total_documents as u64)
     }
 
-        Ok(output)
+    #[cfg(feature = "enterprise")] // only used in enterprise edition for now
+    pub(super) fn export_no_index(
+        &self,
+        target: TargetInstance<'_>,
+        export_old_remote_name: &str,
+        network_change_origin: &Origin,
+        agent: &ureq::Agent,
+        must_stop_processing: &MustStopProcessing,
+    ) -> Result<(), Error> {
+        let bearer = target.api_key.map(|api_key| format!("Bearer {api_key}"));
+        let url = format!("{base_url}/network", base_url = target.base_url,);
+
+        {
+            let _ = handle_response(
+                target.remote_name,
+                retry(must_stop_processing, || {
+                    let request = agent.patch(&url);
+                    let mut request = set_network_ureq_headers(
+                        request,
+                        &ImportData {
+                            remote_name: export_old_remote_name.to_string(),
+                            index_name: None,
+                            document_count: 0,
+                        },
+                        network_change_origin,
+                        &ImportMetadata {
+                            index_count: 0,
+                            task_key: None,
+                            total_index_documents: 0,
+                        },
+                    );
+                    request = request.set("Content-Type", "application/json");
+                    if let Some(bearer) = &bearer {
+                        request = request.set("Authorization", bearer);
+                    }
+                    request
+                        .send_json(
+                            // empty payload that will be disregarded
+                            serde_json::Value::Object(Default::default()),
+                        )
+                        .map_err(into_backoff_error)
+                }),
+            )?;
+        }
+
+        Ok(())
+    }
+}
+
+fn set_network_ureq_headers(
+    request: ureq::Request,
+    import_data: &ImportData,
+    origin: &Origin,
+    metadata: &ImportMetadata,
+) -> ureq::Request {
+    let request = RequestWrapper(request);
+
+    let ImportMetadata { index_count, task_key, total_index_documents } = metadata;
+    let Origin { remote_name: origin_remote, task_uid, network_version } = origin;
+    let ImportData { remote_name: import_remote, index_name, document_count } = import_data;
+
+    let request = request
+        .set_origin_remote(origin_remote)
+        .set_origin_task_uid(*task_uid)
+        .set_origin_network_version(*network_version)
+        .set_import_remote(import_remote)
+        .set_import_docs(*document_count)
+        .set_import_index_count(*index_count)
+        .set_import_index_docs(*total_index_documents);
+
+    let request = if let Some(index_name) = index_name.as_deref() {
+        request.set_import_index(index_name)
+    } else {
+        request
+    };
+    let RequestWrapper(request) = if let Some(task_key) = task_key {
+        request.set_import_task_key(*task_key)
+    } else {
+        request
+    };
+
+    request
+}
+
+struct RequestWrapper(ureq::Request);
+impl headers::SetHeader for RequestWrapper {
+    fn set_header(self, name: &str, value: &str) -> Self {
+        Self(self.0.set(name, value))
+    }
+}
+
+#[allow(clippy::too_many_arguments)]
+fn send_buffer<'a>(
+    buffer: &'a [u8],
+    mut compressed_buffer: &'a mut Vec<u8>,
+    must_stop_processing: &MustStopProcessing,
+    agent: &ureq::Agent,
+    documents_url: &'a str,
+    remote_name: Option<&str>,
+    bearer: Option<&'a str>,
+    task_network: Option<&(ImportData, Origin, ImportMetadata)>,
+    err: &'a impl Fn(milli::Error) -> crate::Error,
+) -> Result<ControlFlow<(), ()>> {
+    // We compress the documents before sending them
+    let mut encoder: GzEncoder<&mut &mut Vec<u8>> =
+        GzEncoder::new(&mut compressed_buffer, Compression::default());
+    encoder.write_all(buffer).map_err(milli::Error::from).map_err(err)?;
+    encoder.finish().map_err(milli::Error::from).map_err(err)?;
+
+    let res = retry(must_stop_processing, || {
+        let mut request = agent.post(documents_url);
+        request = request.set("Content-Type", "application/x-ndjson");
+        request = request.set("Content-Encoding", "gzip");
+        if let Some(bearer) = bearer {
+            request = request.set("Authorization", bearer);
+        }
+        if let Some((import_data, origin, metadata)) = task_network {
+            request = set_network_ureq_headers(request, import_data, origin, metadata);
+        }
+        request.send_bytes(compressed_buffer).map_err(into_backoff_error)
+    });
+
+    handle_response(remote_name, res)
+}
+
+fn handle_response(remote_name: Option<&str>, res: Result<Response>) -> Result<ControlFlow<()>> {
+    let remote_name = remote_name.unwrap_or("unnamed");
+    match res {
+        Ok(_response) => Ok(ControlFlow::Continue(())),
+        Err(Error::FromRemoteWhenExporting { code, .. })
+            if code == Code::ImportTaskAlreadyReceived.name() =>
+        {
+            Ok(ControlFlow::Continue(()))
+        }
+        Err(Error::FromRemoteWhenExporting { code, message, .. })
+            if code == Code::ImportTaskUnknownRemote.name() =>
+        {
+            tracing::warn!("remote `{remote_name}` answered with: {message}");
+            Ok(ControlFlow::Break(()))
+        }
+        // note: there has already been many attempts to get this due to exponential backoff
+        Err(Error::FromRemoteWhenExporting { code, message, .. })
+            if code == Code::ImportTaskWithoutNetworkTask.name() =>
+        {
+            tracing::warn!("remote `{remote_name}` answered with: {message}");
+            Ok(ControlFlow::Break(()))
+        }
+        Err(e) => {
+            tracing::warn!("error while exporting: {e}");
+            Err(e)
+        }
     }
 }
 
@@ -374,4 +593,65 @@ fn ureq_error_into_error(error: ureq::Error) -> Error {
     }
 }
 
+// export_one_index arguments
+pub(super) struct TargetInstance<'a> {
+    pub(super) remote_name: Option<&'a str>,
+    pub(super) base_url: &'a str,
+    pub(super) api_key: Option<&'a str>,
+}
+
+pub(super) struct ExportOptions<'a> {
+    pub(super) index_uid: &'a str,
+    pub(super) payload_size: Option<&'a Byte>,
+    pub(super) override_settings: bool,
+    pub(super) export_mode: ExportMode<'a>,
+}
+
+impl ExportOptions<'_> {
+    fn task_network(
+        &self,
+        total_index_documents: u64,
+    ) -> Option<(ImportData, Origin, ImportMetadata)> {
+        if let ExportMode::NetworkBalancing {
+            index_count,
+            export_old_remote_name,
+            network_change_origin,
+        } = self.export_mode
+        {
+            Some((
+                ImportData {
+                    remote_name: export_old_remote_name.to_string(),
+                    index_name: Some(self.index_uid.to_string()),
+                    document_count: 0,
+                },
+                network_change_origin.clone(),
+                ImportMetadata { index_count, task_key: None, total_index_documents },
+            ))
+        } else {
+            None
+        }
+    }
+}
+
+pub(super) struct ExportContext<'a> {
+    pub(super) index: &'a meilisearch_types::milli::Index,
+    pub(super) index_rtxn: &'a milli::heed::RoTxn<'a>,
+    pub(super) universe: &'a RoaringBitmap,
+    pub(super) progress: &'a Progress,
+    pub(super) agent: &'a ureq::Agent,
+    pub(super) must_stop_processing: &'a MustStopProcessing,
+}
+
+pub(super) enum ExportMode<'a> {
+    ExportRoute,
+    #[cfg_attr(not(feature = "enterprise"), allow(dead_code))]
+    NetworkBalancing {
+        index_count: u64,
+
+        export_old_remote_name: &'a str,
+        network_change_origin: &'a Origin,
+    },
+}
+
+// progress related
 enum ExportIndex {}

crates/index-scheduler/src/scheduler/process_index_operation.rs

@@ -8,6 +8,7 @@ use meilisearch_types::milli::progress::{EmbedderStats, Progress};
 use meilisearch_types::milli::update::new::indexer::{self, UpdateByFunction};
 use meilisearch_types::milli::update::DocumentAdditionResult;
 use meilisearch_types::milli::{self, ChannelCongestion, Filter};
+use meilisearch_types::network::Network;
 use meilisearch_types::settings::apply_settings_to_builder;
 use meilisearch_types::tasks::{Details, KindWithContent, Status, Task};
 use meilisearch_types::Index;
@@ -36,6 +37,7 @@ impl IndexScheduler {
         operation: IndexOperation,
         progress: &Progress,
         embedder_stats: Arc<EmbedderStats>,
+        network: &Network,
     ) -> Result<(Vec<Task>, Option<ChannelCongestion>)> {
         let indexer_alloc = Bump::new();
         let started_processing_at = std::time::Instant::now();
@@ -67,8 +69,6 @@ impl IndexScheduler {
             IndexOperation::DocumentOperation { index_uid, primary_key, operations, mut tasks } => {
                 progress.update_progress(DocumentOperationProgress::RetrievingConfig);
 
-                let network = self.network();
-
                 let shards = network.shards();
 
                 // TODO: at some point, for better efficiency we might want to reuse the bumpalo for successive batches.
@@ -77,8 +77,8 @@ impl IndexScheduler {
                 let mut content_files = Vec::new();
                 for operation in &operations {
                     match operation {
-                        DocumentOperation::Replace(content_uuid)
+                        DocumentOperation::Replace { content_file: content_uuid, .. }
-                        | DocumentOperation::Update(content_uuid) => {
+                        | DocumentOperation::Update { content_file: content_uuid, .. } => {
                             let content_file = self.queue.file_store.get_update(*content_uuid)?;
                             let mmap = unsafe { memmap2::Mmap::map(&content_file)? };
                             content_files.push(mmap);
@@ -100,16 +100,16 @@ impl IndexScheduler {
                 let embedders = self.embedders(index_uid.clone(), embedders)?;
                 for operation in operations {
                     match operation {
-                        DocumentOperation::Replace(_content_uuid) => {
+                        DocumentOperation::Replace { content_file: _, on_missing_document } => {
                             let mmap = content_files_iter.next().unwrap();
                             indexer
-                                .replace_documents(mmap)
+                                .replace_documents(mmap, on_missing_document)
                                 .map_err(|e| Error::from_milli(e, Some(index_uid.clone())))?;
                         }
-                        DocumentOperation::Update(_content_uuid) => {
+                        DocumentOperation::Update { content_file: _, on_missing_document } => {
                             let mmap = content_files_iter.next().unwrap();
                             indexer
-                                .update_documents(mmap)
+                                .update_documents(mmap, on_missing_document)
                                 .map_err(|e| Error::from_milli(e, Some(index_uid.clone())))?;
                         }
                         DocumentOperation::Delete(document_ids) => {
@@ -504,6 +504,7 @@ impl IndexScheduler {
                     },
                     progress,
                     embedder_stats.clone(),
+                    network,
                 )?;
 
                 let (settings_tasks, _congestion) = self.apply_index_operation(
@@ -512,6 +513,7 @@ impl IndexScheduler {
                     IndexOperation::Settings { index_uid, settings, tasks: settings_tasks },
                     progress,
                     embedder_stats,
+                    network,
                 )?;
 
                 let mut tasks = settings_tasks;

crates/index-scheduler/src/scheduler/process_snapshot_creation.rs

@@ -12,7 +12,7 @@ use crate::processing::{AtomicUpdateFileStep, SnapshotCreationProgress};
 use crate::queue::TaskQueue;
 use crate::{Error, IndexScheduler, Result};
 
-const UPDATE_FILES_DIR_NAME: &str = "update_files";
+pub(crate) const UPDATE_FILES_DIR_NAME: &str = "update_files";
 
 /// # Safety
 ///
@@ -230,407 +230,4 @@ impl IndexScheduler {
 
         Ok(tasks)
     }
-
-    #[cfg(unix)]
-    pub(super) async fn process_snapshot_to_s3(
-        &self,
-        progress: Progress,
-        opts: meilisearch_types::milli::update::S3SnapshotOptions,
-        mut tasks: Vec<Task>,
-    ) -> Result<Vec<Task>> {
-        use meilisearch_types::milli::update::S3SnapshotOptions;
-
-        let S3SnapshotOptions {
-            s3_bucket_url,
-            s3_bucket_region,
-            s3_bucket_name,
-            s3_snapshot_prefix,
-            s3_access_key,
-            s3_secret_key,
-            s3_max_in_flight_parts,
-            s3_compression_level: level,
-            s3_signature_duration,
-            s3_multipart_part_size,
-        } = opts;
-
-        let must_stop_processing = self.scheduler.must_stop_processing.clone();
-        let retry_backoff = backoff::ExponentialBackoff::default();
-        let db_name = {
-            let mut base_path = self.env.path().to_owned();
-            base_path.pop();
-            base_path.file_name().and_then(OsStr::to_str).unwrap_or("data.ms").to_string()
-        };
-
-        let (reader, writer) = std::io::pipe()?;
-        let uploader_task = tokio::spawn(multipart_stream_to_s3(
-            s3_bucket_url,
-            s3_bucket_region,
-            s3_bucket_name,
-            s3_snapshot_prefix,
-            s3_access_key,
-            s3_secret_key,
-            s3_max_in_flight_parts,
-            s3_signature_duration,
-            s3_multipart_part_size,
-            must_stop_processing,
-            retry_backoff,
-            db_name,
-            reader,
-        ));
-
-        let index_scheduler = IndexScheduler::private_clone(self);
-        let builder_task = tokio::task::spawn_blocking(move || {
-            stream_tarball_into_pipe(progress, level, writer, index_scheduler)
-        });
-
-        let (uploader_result, builder_result) = tokio::join!(uploader_task, builder_task);
-
-        // Check uploader result first to early return on task abortion.
-        // safety: JoinHandle can return an error if the task was aborted, cancelled, or panicked.
-        uploader_result.unwrap()?;
-        builder_result.unwrap()?;
-
-        for task in &mut tasks {
-            task.status = Status::Succeeded;
-        }
-
-        Ok(tasks)
-    }
-}
-
-/// Streams a tarball of the database content into a pipe.
-#[cfg(unix)]
-fn stream_tarball_into_pipe(
-    progress: Progress,
-    level: u32,
-    writer: std::io::PipeWriter,
-    index_scheduler: IndexScheduler,
-) -> std::result::Result<(), Error> {
-    use std::io::Write as _;
-    use std::path::Path;
-
-    let writer = flate2::write::GzEncoder::new(writer, flate2::Compression::new(level));
-    let mut tarball = tar::Builder::new(writer);
-
-    // 1. Snapshot the version file
-    tarball
-        .append_path_with_name(&index_scheduler.scheduler.version_file_path, VERSION_FILE_NAME)?;
-
-    // 2. Snapshot the index scheduler LMDB env
-    progress.update_progress(SnapshotCreationProgress::SnapshotTheIndexScheduler);
-    let tasks_env_file = index_scheduler.env.try_clone_inner_file()?;
-    let path = Path::new("tasks").join("data.mdb");
-    append_file_to_tarball(&mut tarball, path, tasks_env_file)?;
-
-    // 2.3 Create a read transaction on the index-scheduler
-    let rtxn = index_scheduler.env.read_txn()?;
-
-    // 2.4 Create the update files directory
-    //     And only copy the update files of the enqueued tasks
-    progress.update_progress(SnapshotCreationProgress::SnapshotTheUpdateFiles);
-    let enqueued = index_scheduler.queue.tasks.get_status(&rtxn, Status::Enqueued)?;
-    let (atomic, update_file_progress) = AtomicUpdateFileStep::new(enqueued.len() as u32);
-    progress.update_progress(update_file_progress);
-
-    // We create the update_files directory so that it
-    // always exists even if there are no update files
-    let update_files_dir = Path::new(UPDATE_FILES_DIR_NAME);
-    let src_update_files_dir = {
-        let mut path = index_scheduler.env.path().to_path_buf();
-        path.pop();
-        path.join(UPDATE_FILES_DIR_NAME)
-    };
-    tarball.append_dir(update_files_dir, src_update_files_dir)?;
-
-    for task_id in enqueued {
-        let task = index_scheduler
-            .queue
-            .tasks
-            .get_task(&rtxn, task_id)?
-            .ok_or(Error::CorruptedTaskQueue)?;
-        if let Some(content_uuid) = task.content_uuid() {
-            use std::fs::File;
-
-            let src = index_scheduler.queue.file_store.update_path(content_uuid);
-            let mut update_file = File::open(src)?;
-            let path = update_files_dir.join(content_uuid.to_string());
-            tarball.append_file(path, &mut update_file)?;
-        }
-        atomic.fetch_add(1, Ordering::Relaxed);
-    }
-
-    // 3. Snapshot every indexes
-    progress.update_progress(SnapshotCreationProgress::SnapshotTheIndexes);
-    let index_mapping = index_scheduler.index_mapper.index_mapping;
-    let nb_indexes = index_mapping.len(&rtxn)? as u32;
-    let indexes_dir = Path::new("indexes");
-    let indexes_references: Vec<_> = index_scheduler
-        .index_mapper
-        .index_mapping
-        .iter(&rtxn)?
-        .map(|res| res.map_err(Error::from).map(|(name, uuid)| (name.to_string(), uuid)))
-        .collect::<Result<_, Error>>()?;
-
-    // It's prettier to use a for loop instead of the IndexMapper::try_for_each_index
-    // method, especially when we need to access the UUID, local path and index number.
-    for (i, (name, uuid)) in indexes_references.into_iter().enumerate() {
-        progress.update_progress(VariableNameStep::<SnapshotCreationProgress>::new(
-            &name, i as u32, nb_indexes,
-        ));
-        let path = indexes_dir.join(uuid.to_string()).join("data.mdb");
-        let index = index_scheduler.index_mapper.index(&rtxn, &name)?;
-        let index_file = index.try_clone_inner_file()?;
-        tracing::trace!("Appending index file for {name} in {}", path.display());
-        append_file_to_tarball(&mut tarball, path, index_file)?;
-    }
-
-    drop(rtxn);
-
-    // 4. Snapshot the auth LMDB env
-    progress.update_progress(SnapshotCreationProgress::SnapshotTheApiKeys);
-    let auth_env_file = index_scheduler.scheduler.auth_env.try_clone_inner_file()?;
-    let path = Path::new("auth").join("data.mdb");
-    append_file_to_tarball(&mut tarball, path, auth_env_file)?;
-
-    let mut gzencoder = tarball.into_inner()?;
-    gzencoder.flush()?;
-    gzencoder.try_finish()?;
-    let mut writer = gzencoder.finish()?;
-    writer.flush()?;
-
-    Result::<_, Error>::Ok(())
-}
-
-#[cfg(unix)]
-fn append_file_to_tarball<W, P>(
-    tarball: &mut tar::Builder<W>,
-    path: P,
-    mut auth_env_file: fs::File,
-) -> Result<(), Error>
-where
-    W: std::io::Write,
-    P: AsRef<std::path::Path>,
-{
-    use std::io::{Seek as _, SeekFrom};
-
-    // Note: A previous snapshot operation may have left the cursor
-    //       at the end of the file so we need to seek to the start.
-    auth_env_file.seek(SeekFrom::Start(0))?;
-    tarball.append_file(path, &mut auth_env_file)?;
-    Ok(())
-}
-
-/// Streams the content read from the given reader to S3.
-#[cfg(unix)]
-#[allow(clippy::too_many_arguments)]
-async fn multipart_stream_to_s3(
-    s3_bucket_url: String,
-    s3_bucket_region: String,
-    s3_bucket_name: String,
-    s3_snapshot_prefix: String,
-    s3_access_key: String,
-    s3_secret_key: String,
-    s3_max_in_flight_parts: std::num::NonZero<usize>,
-    s3_signature_duration: std::time::Duration,
-    s3_multipart_part_size: u64,
-    must_stop_processing: super::MustStopProcessing,
-    retry_backoff: backoff::exponential::ExponentialBackoff<backoff::SystemClock>,
-    db_name: String,
-    reader: std::io::PipeReader,
-) -> Result<(), Error> {
-    use std::collections::VecDeque;
-    use std::io;
-    use std::os::fd::OwnedFd;
-    use std::path::PathBuf;
-
-    use bytes::{Bytes, BytesMut};
-    use reqwest::{Client, Response};
-    use rusty_s3::actions::CreateMultipartUpload;
-    use rusty_s3::{Bucket, BucketError, Credentials, S3Action as _, UrlStyle};
-    use tokio::task::JoinHandle;
-
-    let reader = OwnedFd::from(reader);
-    let reader = tokio::net::unix::pipe::Receiver::from_owned_fd(reader)?;
-    let s3_snapshot_prefix = PathBuf::from(s3_snapshot_prefix);
-    let url =
-        s3_bucket_url.parse().map_err(BucketError::ParseError).map_err(Error::S3BucketError)?;
-    let bucket = Bucket::new(url, UrlStyle::Path, s3_bucket_name, s3_bucket_region)
-        .map_err(Error::S3BucketError)?;
-    let credential = Credentials::new(s3_access_key, s3_secret_key);
-
-    // Note for the future (rust 1.91+): use with_added_extension, it's prettier
-    let object_path = s3_snapshot_prefix.join(format!("{db_name}.snapshot"));
-    // Note: It doesn't work on Windows and if a port to this platform is needed,
-    //       use the slash-path crate or similar to get the correct path separator.
-    let object = object_path.display().to_string();
-
-    let action = bucket.create_multipart_upload(Some(&credential), &object);
-    let url = action.sign(s3_signature_duration);
-
-    let client = Client::new();
-    let resp = client.post(url).send().await.map_err(Error::S3HttpError)?;
-    let status = resp.status();
-
-    let body = match resp.error_for_status_ref() {
-        Ok(_) => resp.text().await.map_err(Error::S3HttpError)?,
-        Err(_) => {
-            return Err(Error::S3Error { status, body: resp.text().await.unwrap_or_default() })
-        }
-    };
-
-    let multipart =
-        CreateMultipartUpload::parse_response(&body).map_err(|e| Error::S3XmlError(Box::new(e)))?;
-    tracing::debug!("Starting the upload of the snapshot to {object}");
-
-    // We use this bumpalo for etags strings.
-    let bump = bumpalo::Bump::new();
-    let mut etags = Vec::<&str>::new();
-    let mut in_flight = VecDeque::<(JoinHandle<reqwest::Result<Response>>, Bytes)>::with_capacity(
-        s3_max_in_flight_parts.get(),
-    );
-
-    // Part numbers start at 1 and cannot be larger than 10k
-    for part_number in 1u16.. {
-        if must_stop_processing.get() {
-            return Err(Error::AbortedTask);
-        }
-
-        let part_upload =
-            bucket.upload_part(Some(&credential), &object, part_number, multipart.upload_id());
-        let url = part_upload.sign(s3_signature_duration);
-
-        // Wait for a buffer to be ready if there are in-flight parts that landed
-        let mut buffer = if in_flight.len() >= s3_max_in_flight_parts.get() {
-            let (handle, buffer) = in_flight.pop_front().expect("At least one in flight request");
-            let resp = join_and_map_error(handle).await?;
-            extract_and_append_etag(&bump, &mut etags, resp.headers())?;
-
-            let mut buffer = match buffer.try_into_mut() {
-                Ok(buffer) => buffer,
-                Err(_) => unreachable!("All bytes references were consumed in the task"),
-            };
-            buffer.clear();
-            buffer
-        } else {
-            BytesMut::with_capacity(s3_multipart_part_size as usize)
-        };
-
-        // If we successfully read enough bytes,
-        // we can continue and send the buffer/part
-        while buffer.len() < (s3_multipart_part_size as usize / 2) {
-            // Wait for the pipe to be readable
-
-            reader.readable().await?;
-
-            match reader.try_read_buf(&mut buffer) {
-                Ok(0) => break,
-                // We read some bytes but maybe not enough
-                Ok(_) => continue,
-                // The readiness event is a false positive.
-                Err(ref e) if e.kind() == io::ErrorKind::WouldBlock => continue,
-                Err(e) => return Err(e.into()),
-            }
-        }
-
-        if buffer.is_empty() {
-            // Break the loop if the buffer is
-            // empty after we tried to read bytes
-            break;
-        }
-
-        let body = buffer.freeze();
-        tracing::trace!("Sending part {part_number}");
-        let task = tokio::spawn({
-            let client = client.clone();
-            let body = body.clone();
-            backoff::future::retry(retry_backoff.clone(), move || {
-                let client = client.clone();
-                let url = url.clone();
-                let body = body.clone();
-                async move {
-                    match client.put(url).body(body).send().await {
-                        Ok(resp) if resp.status().is_client_error() => {
-                            resp.error_for_status().map_err(backoff::Error::Permanent)
-                        }
-                        Ok(resp) => Ok(resp),
-                        Err(e) => Err(backoff::Error::transient(e)),
-                    }
-                }
-            })
-        });
-        in_flight.push_back((task, body));
-    }
-
-    for (handle, _buffer) in in_flight {
-        let resp = join_and_map_error(handle).await?;
-        extract_and_append_etag(&bump, &mut etags, resp.headers())?;
-    }
-
-    tracing::debug!("Finalizing the multipart upload");
-
-    let action = bucket.complete_multipart_upload(
-        Some(&credential),
-        &object,
-        multipart.upload_id(),
-        etags.iter().map(AsRef::as_ref),
-    );
-    let url = action.sign(s3_signature_duration);
-    let body = action.body();
-    let resp = backoff::future::retry(retry_backoff, move || {
-        let client = client.clone();
-        let url = url.clone();
-        let body = body.clone();
-        async move {
-            match client.post(url).body(body).send().await {
-                Ok(resp) if resp.status().is_client_error() => {
-                    Err(backoff::Error::Permanent(Error::S3Error {
-                        status: resp.status(),
-                        body: resp.text().await.unwrap_or_default(),
-                    }))
-                }
-                Ok(resp) => Ok(resp),
-                Err(e) => Err(backoff::Error::transient(Error::S3HttpError(e))),
-            }
-        }
-    })
-    .await?;
-
-    let status = resp.status();
-    let body = resp.text().await.map_err(|e| Error::S3Error { status, body: e.to_string() })?;
-    if status.is_success() {
-        Ok(())
-    } else {
-        Err(Error::S3Error { status, body })
-    }
-}
-
-#[cfg(unix)]
-async fn join_and_map_error(
-    join_handle: tokio::task::JoinHandle<Result<reqwest::Response, reqwest::Error>>,
-) -> Result<reqwest::Response> {
-    // safety: Panic happens if the task (JoinHandle) was aborted, cancelled, or panicked
-    let request = join_handle.await.unwrap();
-    let resp = request.map_err(Error::S3HttpError)?;
-    match resp.error_for_status_ref() {
-        Ok(_) => Ok(resp),
-        Err(_) => Err(Error::S3Error {
-            status: resp.status(),
-            body: resp.text().await.unwrap_or_default(),
-        }),
-    }
-}
-
-#[cfg(unix)]
-fn extract_and_append_etag<'b>(
-    bump: &'b bumpalo::Bump,
-    etags: &mut Vec<&'b str>,
-    headers: &reqwest::header::HeaderMap,
-) -> Result<()> {
-    use reqwest::header::ETAG;
-
-    let etag = headers.get(ETAG).ok_or_else(|| Error::S3XmlError("Missing ETag header".into()))?;
-    let etag = etag.to_str().map_err(|e| Error::S3XmlError(Box::new(e)))?;
-    etags.push(bump.alloc_str(etag));
-
-    Ok(())
 }

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_enqueued_task/cancel_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: canceled, canceled_by: 1, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: canceled, canceled_by: 1, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, batch_uid: 0, status: succeeded, details: { matched_tasks: 1, canceled_tasks: Some(1), original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_enqueued_task/initial_tasks_enqueued.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, status: enqueued, details: { matched_tasks: 1, canceled_tasks: None, original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_mix_of_tasks/aborted_indexation.snap

@@ -7,9 +7,9 @@ source: crates/index-scheduler/src/scheduler/test.rs
 {uid: 1, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"beavero":1}}, stop reason: "batched all enqueued tasks for index `beavero`", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, status: enqueued, details: { matched_tasks: 3, canceled_tasks: None, original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0, 1, 2]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_mix_of_tasks/cancel_processed.snap

@@ -6,9 +6,9 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, batch_uid: 1, status: canceled, canceled_by: 3, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, batch_uid: 1, status: canceled, canceled_by: 3, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, batch_uid: 1, status: canceled, canceled_by: 3, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, batch_uid: 1, status: canceled, canceled_by: 3, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, batch_uid: 1, status: succeeded, details: { matched_tasks: 3, canceled_tasks: Some(2), original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0, 1, 2]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_mix_of_tasks/first_task_processed.snap

@@ -6,9 +6,9 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [1,2,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_mix_of_tasks/processing_second_task_cancel_enqueued.snap

@@ -7,9 +7,9 @@ source: crates/index-scheduler/src/scheduler/test.rs
 {uid: 1, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"beavero":1}}, stop reason: "batched all enqueued tasks for index `beavero`", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "beavero", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "wolfo", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000002, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, status: enqueued, details: { matched_tasks: 3, canceled_tasks: None, original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0, 1, 2]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_processing_task/aborted_indexation.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 {uid: 0, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"catto":1}}, stop reason: "batched all enqueued tasks", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, status: enqueued, details: { matched_tasks: 1, canceled_tasks: None, original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_processing_task/cancel_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: canceled, canceled_by: 1, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: canceled, canceled_by: 1, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, batch_uid: 0, status: succeeded, details: { matched_tasks: 1, canceled_tasks: Some(1), original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_processing_task/cancel_task_registered.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 {uid: 0, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"catto":1}}, stop reason: "batched all enqueued tasks", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, status: enqueued, details: { matched_tasks: 1, canceled_tasks: None, original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_processing_task/initial_task_processing.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 {uid: 0, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"catto":1}}, stop reason: "batched all enqueued tasks", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_processing_task/registered_the_first_task.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_succeeded_task/cancel_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, batch_uid: 1, status: succeeded, details: { matched_tasks: 1, canceled_tasks: Some(0), original_filter: "test_query" }, kind: TaskCancelation { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_succeeded_task/initial_task_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued []

crates/index-scheduler/src/scheduler/snapshots/test.rs/cancel_succeeded_task/registered_the_first_task.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion/before_index_creation.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, batch_uid: 0, status: succeeded, details: { primary_key: None, old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "doggos", primary_key: None }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, status: enqueued, details: { deleted_documents: None }, kind: IndexDeletion { index_uid: "doggos" }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion/both_task_succeeded.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, batch_uid: 0, status: succeeded, details: { primary_key: None, old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "doggos", primary_key: None }}
-1 {uid: 1, batch_uid: 1, status: succeeded, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, batch_uid: 1, status: succeeded, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, batch_uid: 1, status: succeeded, details: { deleted_documents: Some(0) }, kind: IndexDeletion { index_uid: "doggos" }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion/registered_the_second_task.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: None, old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "doggos", primary_key: None }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,1,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion/registered_the_third_task.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: None, old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "doggos", primary_key: None }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, status: enqueued, details: { deleted_documents: None }, kind: IndexDeletion { index_uid: "doggos" }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion_on_unexisting_index/1.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, status: enqueued, details: { deleted_documents: None }, kind: IndexDeletion { index_uid: "doggos" }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/document_addition_and_index_deletion_on_unexisting_index/2.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(0) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, batch_uid: 0, status: succeeded, details: { deleted_documents: Some(0) }, kind: IndexDeletion { index_uid: "doggos" }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_delete_same_task_twice/initial_tasks_enqueued.snap

@@ -1,14 +1,13 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,1,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_delete_same_task_twice/initial_tasks_processed.snap

@@ -6,8 +6,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [1,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_delete_same_task_twice/task_deletion_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, batch_uid: 1, status: succeeded, details: { matched_tasks: 1, deleted_tasks: Some(1), original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0]> }}
 3 {uid: 3, batch_uid: 1, status: succeeded, details: { matched_tasks: 1, deleted_tasks: Some(0), original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_deleteable/after_registering_the_task_deletion.snap

@@ -6,8 +6,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, status: enqueued, details: { matched_tasks: 1, deleted_tasks: None, original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_deleteable/initial_tasks_enqueued.snap

@@ -1,14 +1,13 @@
 ---
 source: crates/index-scheduler/src/scheduler/test.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,1,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_deleteable/initial_tasks_processed.snap

@@ -6,8 +6,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [1,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_deleteable/task_deletion_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 2 {uid: 2, batch_uid: 1, status: succeeded, details: { matched_tasks: 1, deleted_tasks: Some(1), original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_undeleteable/initial_tasks_enqueued.snap

@@ -7,8 +7,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: Some("mouse"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "catto", primary_key: Some("mouse") }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,1,2,]

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_undeleteable/task_deletion_done.snap

@@ -7,8 +7,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: Some("mouse"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "catto", primary_key: Some("mouse") }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, batch_uid: 0, status: succeeded, details: { matched_tasks: 2, deleted_tasks: Some(0), original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0, 1]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_undeleteable/task_deletion_enqueued.snap

@@ -7,8 +7,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: Some("mouse"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "catto", primary_key: Some("mouse") }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, status: enqueued, details: { matched_tasks: 2, deleted_tasks: None, original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0, 1]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test.rs/task_deletion_undeleteable/task_deletion_processing.snap

@@ -8,8 +8,8 @@ source: crates/index-scheduler/src/scheduler/test.rs
 ----------------------------------------------------------------------
 ### All Tasks:
 0 {uid: 0, status: enqueued, details: { primary_key: Some("mouse"), old_new_uid: None, new_index_uid: None }, kind: IndexCreation { index_uid: "catto", primary_key: Some("mouse") }}
-1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+1 {uid: 1, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "catto", primary_key: None, method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
-2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true }}
+2 {uid: 2, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggo", primary_key: Some("bone"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000001, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 3 {uid: 3, status: enqueued, details: { matched_tasks: 2, deleted_tasks: None, original_filter: "test_query" }, kind: TaskDeletion { query: "test_query", tasks: RoaringBitmap<[0, 1]> }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition/after_register.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test_document_addition.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition/after_the_batch_creation.snap

@@ -7,7 +7,7 @@ source: crates/index-scheduler/src/scheduler/test_document_addition.rs
 {uid: 0, details: {"receivedDocuments":1,"indexedDocuments":null}, stats: {"totalNbTasks":1,"status":{"processing":1},"types":{"documentAdditionOrUpdate":1},"indexUids":{"doggos":1}}, stop reason: "batched all enqueued tasks", }
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 1, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition/once_everything_is_processed.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test_document_addition.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 1, indexed_documents: Some(1) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 1, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued []

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition_and_document_deletion/after_processing_the_batch.snap

@@ -6,7 +6,7 @@ source: crates/index-scheduler/src/scheduler/test_document_addition.rs
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 3, indexed_documents: Some(3) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true }}
+0 {uid: 0, batch_uid: 0, status: succeeded, details: { received_documents: 3, indexed_documents: Some(3) }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, batch_uid: 0, status: succeeded, details: { received_document_ids: 2, deleted_documents: Some(2) }, kind: DocumentDeletion { index_uid: "doggos", documents_ids: ["1", "2"] }}
 ----------------------------------------------------------------------
 ### Status:

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition_and_document_deletion/registered_the_first_task.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test_document_addition.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 3, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 3, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true, on_missing_document: Create }}
 ----------------------------------------------------------------------
 ### Status:
 enqueued [0,]

crates/index-scheduler/src/scheduler/snapshots/test_document_addition.rs/document_addition_and_document_deletion/registered_the_second_task.snap

@@ -1,13 +1,12 @@
 ---
 source: crates/index-scheduler/src/scheduler/test_document_addition.rs
-snapshot_kind: text
 ---
 ### Autobatching Enabled = true
 ### Processing batch None:
 []
 ----------------------------------------------------------------------
 ### All Tasks:
-0 {uid: 0, status: enqueued, details: { received_documents: 3, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true }}
+0 {uid: 0, status: enqueued, details: { received_documents: 3, indexed_documents: None }, kind: DocumentAdditionOrUpdate { index_uid: "doggos", primary_key: Some("id"), method: ReplaceDocuments, content_file: 00000000-0000-0000-0000-000000000000, documents_count: 3, allow_index_creation: true, on_missing_document: Create }}
 1 {uid: 1, status: enqueued, details: { received_document_ids: 2, deleted_documents: None }, kind: DocumentDeletion { index_uid: "doggos", documents_ids: ["1", "2"] }}
 ----------------------------------------------------------------------
 ### Status: