📝 document queued sink methods

docs/api/index.md

@@ -135,6 +135,10 @@ BitLogger API navigation.
 - [buffered-sink-flush.md](./buffered-sink-flush.md)
 - [queued-sink.md](./queued-sink.md)
 - [queued-sink-type.md](./queued-sink-type.md)
+- [queued-sink-pending-count.md](./queued-sink-pending-count.md)
+- [queued-sink-dropped-count.md](./queued-sink-dropped-count.md)
+- [queued-sink-drain.md](./queued-sink-drain.md)
+- [queued-sink-flush.md](./queued-sink-flush.md)
 - [queue-overflow-policy.md](./queue-overflow-policy.md)
 - [filter-sink.md](./filter-sink.md)
 - [filter-sink-type.md](./filter-sink-type.md)

docs/api/queued-sink-drain.md

@@ -0,0 +1,75 @@
+---
+name: queued-sink-drain
+group: api
+category: sink
+update-time: 20260613
+description: Drain queued records from a QueuedSink with an optional item limit.
+key-word:
+    - sink
+    - queue
+    - drain
+    - public
+---
+
+## Queued-sink-drain
+
+Drain queued records from a `QueuedSink[S]`. This is the direct sink-level queue delivery API when code owns a queued sink and wants controlled progress into the wrapped sink.
+
+### Interface
+
+```moonbit
+pub fn[S : Sink] QueuedSink::drain(self : QueuedSink[S], max_items~ : Int = -1) -> Int {
+```
+
+#### input
+
+- `self : QueuedSink[S]` - Queued sink whose pending records should be drained to the wrapped sink.
+- `max_items : Int` - Optional upper bound on drained queued records. Negative values mean no explicit bound.
+
+#### output
+
+- `Int` - Number of drained records.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- `max_items == 0` returns `0` immediately.
+- Negative `max_items` means drain up to the current `pending_count()`.
+- The method repeatedly pops from the queue and writes each record to the wrapped sink in order until the limit is reached or the queue becomes empty.
+- The return value is the number of records actually drained.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need Bounded Queue Progress
+
+When queued work should be advanced in chunks:
+```moonbit
+let drained = sink.drain(max_items=16)
+```
+
+In this example, callers cap how much queue backlog is forwarded in one step.
+
+#### When Need Full Manual Drain
+
+When a queued sink should be emptied explicitly:
+```moonbit
+ignore(sink.drain())
+```
+
+In this example, the queue is drained without an explicit item cap.
+
+### Error Case
+
+e.g.:
+- If the queue is already empty, the method returns `0`.
+
+- If callers only need the shorthand that drains the whole queue, `flush()` is the simpler API.
+
+### Notes
+
+1. Prefer this helper when queue progress should be bounded or directly observed.
+
+2. Pair it with `pending_count()` to inspect remaining backlog after a drain step.

docs/api/queued-sink-dropped-count.md

@@ -0,0 +1,76 @@
+---
+name: queued-sink-dropped-count
+group: api
+category: sink
+update-time: 20260613
+description: Read the cumulative dropped-record count from a QueuedSink.
+key-word:
+    - sink
+    - queue
+    - dropped
+    - public
+---
+
+## Queued-sink-dropped-count
+
+Read the cumulative dropped-record count from a `QueuedSink[S]`. This helper exposes the direct sink-level loss metric when a bounded queue overflows.
+
+### Interface
+
+```moonbit
+pub fn[S] QueuedSink::dropped_count(self : QueuedSink[S]) -> Int {
+```
+
+#### input
+
+- `self : QueuedSink[S]` - Queued sink whose cumulative dropped-record metric should be inspected.
+
+#### output
+
+- `Int` - Number of dropped records reported by the queue wrapper.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- The counter value comes from `self.dropped_count.val`.
+- It increases when the queue is full and an incoming record cannot be kept without applying overflow handling.
+- The counter is cumulative for the lifetime of the concrete queued sink value.
+- This helper does not explain why drops happened beyond exposing the count.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need Direct Queue Loss Visibility
+
+When a bounded queue may discard records under pressure:
+```moonbit
+if sink.dropped_count() > 0 {
+  println("queue dropped records")
+}
+```
+
+In this example, direct queue loss becomes visible without inspecting internal fields manually.
+
+#### When Compare Overflow Tuning Changes
+
+When queue capacity or overflow policy should be checked operationally:
+```moonbit
+ignore(sink.dropped_count())
+```
+
+In this example, callers can compare the cumulative drop metric across different queue settings.
+
+### Error Case
+
+e.g.:
+- If the queue never overflows, this counter remains `0`.
+
+- If callers need a higher-level wrapper around runtime-configured sinks, `RuntimeSink::dropped_count()` or `ConfiguredLogger::dropped_count()` may fit better.
+
+### Notes
+
+1. This helper reports cumulative loss, not the exact overflow moment.
+
+2. Pair it with `pending_count()`, `drain()`, and queue configuration when investigating pressure.

docs/api/queued-sink-flush.md

@@ -0,0 +1,74 @@
+---
+name: queued-sink-flush
+group: api
+category: sink
+update-time: 20260613
+description: Flush a QueuedSink by draining its queued records into the wrapped sink.
+key-word:
+    - sink
+    - queue
+    - flush
+    - public
+---
+
+## Queued-sink-flush
+
+Flush a `QueuedSink[S]` by draining its queued records into the wrapped sink. This is the direct sink-level shorthand for draining the whole queue.
+
+### Interface
+
+```moonbit
+pub fn[S : Sink] QueuedSink::flush(self : QueuedSink[S]) -> Int {
+```
+
+#### input
+
+- `self : QueuedSink[S]` - Queued sink whose pending records should be fully drained to the wrapped sink.
+
+#### output
+
+- `Int` - Number of drained records.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- This method delegates directly to `self.drain()`.
+- It drains as many queued records as are currently pending.
+- The return value is the number of records forwarded to the wrapped sink.
+- This helper is the sink-level shorthand for a full manual queue drain.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need Full Queue Flush Semantics
+
+When a queued sink should forward all pending records:
+```moonbit
+let flushed = sink.flush()
+```
+
+In this example, callers can observe how many queued records were forwarded.
+
+#### When Need A Simpler Alternative To Explicit Drain Limits
+
+When bounded draining is unnecessary and the whole queue should be released:
+```moonbit
+ignore(sink.flush())
+```
+
+In this example, the queue is drained completely without specifying `max_items`.
+
+### Error Case
+
+e.g.:
+- If the queue is already empty, the method returns `0`.
+
+- If callers need bounded progress instead of a full drain, `drain(max_items=...)` is the better API.
+
+### Notes
+
+1. Use this helper when code owns a `QueuedSink` directly and wants a full manual queue release.
+
+2. This helper returns a count, unlike `BufferedSink::flush()`.

docs/api/queued-sink-pending-count.md

@@ -0,0 +1,75 @@
+---
+name: queued-sink-pending-count
+group: api
+category: sink
+update-time: 20260613
+description: Read the current queued-record count from a QueuedSink.
+key-word:
+    - sink
+    - queue
+    - backlog
+    - public
+---
+
+## Queued-sink-pending-count
+
+Read the current queued-record count from a `QueuedSink[S]`. This is the direct sink-level backlog metric for records that have been queued but not yet drained to the wrapped sink.
+
+### Interface
+
+```moonbit
+pub fn[S] QueuedSink::pending_count(self : QueuedSink[S]) -> Int {
+```
+
+#### input
+
+- `self : QueuedSink[S]` - Queued sink whose current pending backlog should be inspected.
+
+#### output
+
+- `Int` - Current number of queued records.
+
+### Explanation
+
+Detailed rules explaining key parameters and behaviors
+
+- The return value is `self.queue.length()` at the time of the call.
+- This is a point-in-time metric and may change immediately after it is read.
+- It reflects queued records that have not yet been drained or flushed to the wrapped sink.
+- This helper does not mutate the sink.
+
+### How to Use
+
+Here are some specific examples provided.
+
+#### When Need Direct Queue Backlog Visibility
+
+When code is working with a `QueuedSink` value directly and wants to inspect backlog pressure:
+```moonbit
+let pending = sink.pending_count()
+```
+
+In this example, callers can inspect the current queue size without draining it.
+
+#### When Verify Drain Progress
+
+When manual draining should be checked operationally:
+```moonbit
+ignore(sink.drain(max_items=8))
+ignore(sink.pending_count())
+```
+
+In this example, the metric helps verify whether a drain step reduced queued backlog.
+
+### Error Case
+
+e.g.:
+- This helper does not have a normal failure mode; it only reads current queue length.
+
+- If callers only need simple threshold buffering instead of an explicit queue, `BufferedSink::pending_count()` may be the better API.
+
+### Notes
+
+1. Use this helper for direct visibility into queue backlog on `QueuedSink` values.
+
+2. Pair it with `dropped_count()` when investigating queue pressure.