4238: Task queue webhook r=dureuill a=irevoire

# Prototype `prototype-task-queue-webhook-1`

The prototype is available through Docker by using the following command:

```bash
docker run -p 7700:7700 -v $(pwd)/meili_data:/meili_data getmeili/meilisearch:prototype-task-queue-webhook-1
```

# Pull Request

Implements the task queue webhook.

## Related issue
Fixes https://github.com/meilisearch/meilisearch/issues/4236

## What does this PR do?
- Provide a new cli and env var for the webhook, respectively called `--task-webhook-url` and `MEILI_TASK_WEBHOOK_URL`
- Also supports sending the requests with a custom `Authorization` header by specifying the optional `--task-webhook-authorization-header` CLI parameter or `MEILI_TASK_WEBHOOK_AUTHORIZATION_HEADER` env variable.
- Throw an error if the specified URL is invalid
- Every time a batch is processed, send all the finished tasks into the webhook with our public `TaskView` type as a JSON Line GZIPed body.
- Add one test.

## PR checklist

### Before becoming ready to review
- [x] Add a test
- [x] Compress the data we send
- [x] Chunk and stream the data we send
- [x] Remove the unwrap in the index-scheduler when sending the data fails
- [x] The analytics are missing

### Before merging
- [x] Release a prototype



Co-authored-by: Tamo <tamo@meilisearch.com>
Co-authored-by: Clément Renault <clement@meilisearch.com>
This commit is contained in:
meili-bors[bot]
2023-12-21 14:43:46 +00:00
committed by GitHub
13 changed files with 418 additions and 151 deletions

View File

@ -104,6 +104,7 @@ walkdir = "2.3.3"
yaup = "0.2.1"
serde_urlencoded = "0.7.1"
termcolor = "1.2.0"
url = { version = "2.5.0", features = ["serde"] }
[dev-dependencies]
actix-rt = "2.8.0"

View File

@ -264,6 +264,8 @@ struct Infos {
ignore_snapshot_if_db_exists: bool,
http_addr: bool,
http_payload_size_limit: Byte,
task_queue_webhook: bool,
task_webhook_authorization_header: bool,
log_level: String,
max_indexing_memory: MaxMemory,
max_indexing_threads: MaxThreads,
@ -290,6 +292,8 @@ impl From<Opt> for Infos {
http_addr,
master_key: _,
env,
task_webhook_url,
task_webhook_authorization_header,
max_index_size: _,
max_task_db_size: _,
http_payload_size_limit,
@ -343,6 +347,8 @@ impl From<Opt> for Infos {
http_addr: http_addr != default_http_addr(),
http_payload_size_limit,
experimental_max_number_of_batched_tasks,
task_queue_webhook: task_webhook_url.is_some(),
task_webhook_authorization_header: task_webhook_authorization_header.is_some(),
log_level: log_level.to_string(),
max_indexing_memory,
max_indexing_threads,

View File

@ -228,6 +228,8 @@ fn open_or_create_database_unchecked(
indexes_path: opt.db_path.join("indexes"),
snapshots_path: opt.snapshot_dir.clone(),
dumps_path: opt.dump_dir.clone(),
webhook_url: opt.task_webhook_url.as_ref().map(|url| url.to_string()),
webhook_authorization_header: opt.task_webhook_authorization_header.clone(),
task_db_size: opt.max_task_db_size.get_bytes() as usize,
index_base_map_size: opt.max_index_size.get_bytes() as usize,
enable_mdb_writemap: opt.experimental_reduce_indexing_memory_usage,

View File

@ -21,6 +21,7 @@ use rustls::RootCertStore;
use rustls_pemfile::{certs, pkcs8_private_keys, rsa_private_keys};
use serde::{Deserialize, Serialize};
use sysinfo::{RefreshKind, System, SystemExt};
use url::Url;
const POSSIBLE_ENV: [&str; 2] = ["development", "production"];
@ -28,6 +29,8 @@ const MEILI_DB_PATH: &str = "MEILI_DB_PATH";
const MEILI_HTTP_ADDR: &str = "MEILI_HTTP_ADDR";
const MEILI_MASTER_KEY: &str = "MEILI_MASTER_KEY";
const MEILI_ENV: &str = "MEILI_ENV";
const MEILI_TASK_WEBHOOK_URL: &str = "MEILI_TASK_WEBHOOK_URL";
const MEILI_TASK_WEBHOOK_AUTHORIZATION_HEADER: &str = "MEILI_TASK_WEBHOOK_AUTHORIZATION_HEADER";
#[cfg(feature = "analytics")]
const MEILI_NO_ANALYTICS: &str = "MEILI_NO_ANALYTICS";
const MEILI_HTTP_PAYLOAD_SIZE_LIMIT: &str = "MEILI_HTTP_PAYLOAD_SIZE_LIMIT";
@ -156,6 +159,14 @@ pub struct Opt {
#[serde(default = "default_env")]
pub env: String,
/// Called whenever a task finishes so a third party can be notified.
#[clap(long, env = MEILI_TASK_WEBHOOK_URL)]
pub task_webhook_url: Option<Url>,
/// The Authorization header to send on the webhook URL whenever a task finishes so a third party can be notified.
#[clap(long, env = MEILI_TASK_WEBHOOK_AUTHORIZATION_HEADER)]
pub task_webhook_authorization_header: Option<String>,
/// Deactivates Meilisearch's built-in telemetry when provided.
///
/// Meilisearch automatically collects data from all instances that do not opt out using this flag.
@ -375,6 +386,8 @@ impl Opt {
http_addr,
master_key,
env,
task_webhook_url,
task_webhook_authorization_header,
max_index_size: _,
max_task_db_size: _,
http_payload_size_limit,
@ -409,6 +422,16 @@ impl Opt {
export_to_env_if_not_present(MEILI_MASTER_KEY, master_key);
}
export_to_env_if_not_present(MEILI_ENV, env);
if let Some(task_webhook_url) = task_webhook_url {
export_to_env_if_not_present(MEILI_TASK_WEBHOOK_URL, task_webhook_url.to_string());
}
if let Some(task_webhook_authorization_header) = task_webhook_authorization_header {
export_to_env_if_not_present(
MEILI_TASK_WEBHOOK_AUTHORIZATION_HEADER,
task_webhook_authorization_header,
);
}
#[cfg(feature = "analytics")]
{
export_to_env_if_not_present(MEILI_NO_ANALYTICS, no_analytics.to_string());

View File

@ -8,11 +8,9 @@ use meilisearch_types::deserr::DeserrQueryParamError;
use meilisearch_types::error::deserr_codes::*;
use meilisearch_types::error::{InvalidTaskDateError, ResponseError};
use meilisearch_types::index_uid::IndexUid;
use meilisearch_types::settings::{Settings, Unchecked};
use meilisearch_types::star_or::{OptionStarOr, OptionStarOrList};
use meilisearch_types::tasks::{
serialize_duration, Details, IndexSwap, Kind, KindWithContent, Status, Task,
};
use meilisearch_types::task_view::TaskView;
use meilisearch_types::tasks::{Kind, KindWithContent, Status};
use serde::Serialize;
use serde_json::json;
use time::format_description::well_known::Rfc3339;
@ -37,140 +35,6 @@ pub fn configure(cfg: &mut web::ServiceConfig) {
.service(web::resource("/cancel").route(web::post().to(SeqHandler(cancel_tasks))))
.service(web::resource("/{task_id}").route(web::get().to(SeqHandler(get_task))));
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize)]
#[serde(rename_all = "camelCase")]
pub struct TaskView {
pub uid: TaskId,
#[serde(default)]
pub index_uid: Option<String>,
pub status: Status,
#[serde(rename = "type")]
pub kind: Kind,
pub canceled_by: Option<TaskId>,
#[serde(skip_serializing_if = "Option::is_none")]
pub details: Option<DetailsView>,
pub error: Option<ResponseError>,
#[serde(serialize_with = "serialize_duration", default)]
pub duration: Option<Duration>,
#[serde(with = "time::serde::rfc3339")]
pub enqueued_at: OffsetDateTime,
#[serde(with = "time::serde::rfc3339::option", default)]
pub started_at: Option<OffsetDateTime>,
#[serde(with = "time::serde::rfc3339::option", default)]
pub finished_at: Option<OffsetDateTime>,
}
impl TaskView {
pub fn from_task(task: &Task) -> TaskView {
TaskView {
uid: task.uid,
index_uid: task.index_uid().map(ToOwned::to_owned),
status: task.status,
kind: task.kind.as_kind(),
canceled_by: task.canceled_by,
details: task.details.clone().map(DetailsView::from),
error: task.error.clone(),
duration: task.started_at.zip(task.finished_at).map(|(start, end)| end - start),
enqueued_at: task.enqueued_at,
started_at: task.started_at,
finished_at: task.finished_at,
}
}
}
#[derive(Default, Debug, PartialEq, Eq, Clone, Serialize)]
#[serde(rename_all = "camelCase")]
pub struct DetailsView {
#[serde(skip_serializing_if = "Option::is_none")]
pub received_documents: Option<u64>,
#[serde(skip_serializing_if = "Option::is_none")]
pub indexed_documents: Option<Option<u64>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub primary_key: Option<Option<String>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub provided_ids: Option<usize>,
#[serde(skip_serializing_if = "Option::is_none")]
pub deleted_documents: Option<Option<u64>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub matched_tasks: Option<u64>,
#[serde(skip_serializing_if = "Option::is_none")]
pub canceled_tasks: Option<Option<u64>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub deleted_tasks: Option<Option<u64>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub original_filter: Option<Option<String>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub dump_uid: Option<Option<String>>,
#[serde(skip_serializing_if = "Option::is_none")]
#[serde(flatten)]
pub settings: Option<Box<Settings<Unchecked>>>,
#[serde(skip_serializing_if = "Option::is_none")]
pub swaps: Option<Vec<IndexSwap>>,
}
impl From<Details> for DetailsView {
fn from(details: Details) -> Self {
match details {
Details::DocumentAdditionOrUpdate { received_documents, indexed_documents } => {
DetailsView {
received_documents: Some(received_documents),
indexed_documents: Some(indexed_documents),
..DetailsView::default()
}
}
Details::SettingsUpdate { settings } => {
DetailsView { settings: Some(settings), ..DetailsView::default() }
}
Details::IndexInfo { primary_key } => {
DetailsView { primary_key: Some(primary_key), ..DetailsView::default() }
}
Details::DocumentDeletion {
provided_ids: received_document_ids,
deleted_documents,
} => DetailsView {
provided_ids: Some(received_document_ids),
deleted_documents: Some(deleted_documents),
original_filter: Some(None),
..DetailsView::default()
},
Details::DocumentDeletionByFilter { original_filter, deleted_documents } => {
DetailsView {
provided_ids: Some(0),
original_filter: Some(Some(original_filter)),
deleted_documents: Some(deleted_documents),
..DetailsView::default()
}
}
Details::ClearAll { deleted_documents } => {
DetailsView { deleted_documents: Some(deleted_documents), ..DetailsView::default() }
}
Details::TaskCancelation { matched_tasks, canceled_tasks, original_filter } => {
DetailsView {
matched_tasks: Some(matched_tasks),
canceled_tasks: Some(canceled_tasks),
original_filter: Some(Some(original_filter)),
..DetailsView::default()
}
}
Details::TaskDeletion { matched_tasks, deleted_tasks, original_filter } => {
DetailsView {
matched_tasks: Some(matched_tasks),
deleted_tasks: Some(deleted_tasks),
original_filter: Some(Some(original_filter)),
..DetailsView::default()
}
}
Details::Dump { dump_uid } => {
DetailsView { dump_uid: Some(dump_uid), ..DetailsView::default() }
}
Details::IndexSwap { swaps } => {
DetailsView { swaps: Some(swaps), ..Default::default() }
}
}
}
}
#[derive(Debug, Deserr)]
#[deserr(error = DeserrQueryParamError, rename_all = camelCase, deny_unknown_fields)]
pub struct TasksFilterQuery {

View File

@ -1,4 +1,5 @@
mod errors;
mod webhook;
use meili_snap::insta::assert_json_snapshot;
use time::format_description::well_known::Rfc3339;

View File

@ -0,0 +1,123 @@
//! To test the webhook, we need to spawn a new server with a URL listening for
//! post requests. The webhook handle starts a server and forwards all the
//! received requests into a channel for you to handle.
use std::sync::Arc;
use actix_http::body::MessageBody;
use actix_web::dev::{ServiceFactory, ServiceResponse};
use actix_web::web::{Bytes, Data};
use actix_web::{post, App, HttpResponse, HttpServer};
use meili_snap::{json_string, snapshot};
use meilisearch::Opt;
use tokio::sync::mpsc;
use url::Url;
use crate::common::{default_settings, Server};
use crate::json;
#[post("/")]
async fn forward_body(sender: Data<mpsc::UnboundedSender<Vec<u8>>>, body: Bytes) -> HttpResponse {
let body = body.to_vec();
sender.send(body).unwrap();
HttpResponse::Ok().into()
}
fn create_app(
sender: Arc<mpsc::UnboundedSender<Vec<u8>>>,
) -> actix_web::App<
impl ServiceFactory<
actix_web::dev::ServiceRequest,
Config = (),
Response = ServiceResponse<impl MessageBody>,
Error = actix_web::Error,
InitError = (),
>,
> {
App::new().service(forward_body).app_data(Data::from(sender))
}
struct WebhookHandle {
pub server_handle: tokio::task::JoinHandle<Result<(), std::io::Error>>,
pub url: String,
pub receiver: mpsc::UnboundedReceiver<Vec<u8>>,
}
async fn create_webhook_server() -> WebhookHandle {
let mut log_builder = env_logger::Builder::new();
log_builder.parse_filters("info");
log_builder.init();
let (sender, receiver) = mpsc::unbounded_channel();
let sender = Arc::new(sender);
// By listening on the port 0, the system will give us any available port.
let server =
HttpServer::new(move || create_app(sender.clone())).bind(("127.0.0.1", 0)).unwrap();
let (ip, scheme) = server.addrs_with_scheme()[0];
let url = format!("{scheme}://{ip}/");
let server_handle = tokio::spawn(server.run());
WebhookHandle { server_handle, url, receiver }
}
#[actix_web::test]
async fn test_basic_webhook() {
let WebhookHandle { server_handle, url, mut receiver } = create_webhook_server().await;
let db_path = tempfile::tempdir().unwrap();
let server = Server::new_with_options(Opt {
task_webhook_url: Some(Url::parse(&url).unwrap()),
..default_settings(db_path.path())
})
.await
.unwrap();
let index = server.index("tamo");
// May be flaky: we're relying on the fact that while the first document addition is processed, the other
// operations will be received and will be batched together. If it doesn't happen it's not a problem
// the rest of the test won't assume anything about the number of tasks per batch.
for i in 0..5 {
let (_, _status) = index.add_documents(json!({ "id": i, "doggo": "bone" }), None).await;
}
let mut nb_tasks = 0;
while let Some(payload) = receiver.recv().await {
let payload = String::from_utf8(payload).unwrap();
let jsonl = payload.split('\n');
for json in jsonl {
if json.is_empty() {
break; // we reached EOF
}
nb_tasks += 1;
let json: serde_json::Value = serde_json::from_str(json).unwrap();
snapshot!(
json_string!(json, { ".uid" => "[uid]", ".duration" => "[duration]", ".enqueuedAt" => "[date]", ".startedAt" => "[date]", ".finishedAt" => "[date]" }),
@r###"
{
"uid": "[uid]",
"indexUid": "tamo",
"status": "succeeded",
"type": "documentAdditionOrUpdate",
"canceledBy": null,
"details": {
"receivedDocuments": 1,
"indexedDocuments": 1
},
"error": null,
"duration": "[duration]",
"enqueuedAt": "[date]",
"startedAt": "[date]",
"finishedAt": "[date]"
}
"###);
}
if nb_tasks == 5 {
break;
}
}
assert!(nb_tasks == 5, "We should have received the 5 tasks but only received {nb_tasks}");
server_handle.abort();
}