Documents worker-partition-buckets()

And adds it to every destination that has worker-partition-key()

Author: Robert Fekete

content/chapter-destinations/clickhouse/_index.md

@@ -309,6 +309,8 @@ By default, sending data to ClickHouse doesn't propagate the type information of
 
 *Description:* The username used for authentication.
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-destinations/configuring-destinations-http-nonjava/reference-destination-http-nonjava/_index.md

@@ -422,6 +422,8 @@ In case the server on the specified URL returns a redirect request, {{% param "p
 
 {{% include-headless "chunk/option-destination-http-use-system-cert-store.md" %}}
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-destinations/destination-loki/_index.md

@@ -167,6 +167,8 @@ loki(
 
 *Description:* The URL of the Loki endpoint, including the gRPC listen port of your Loki deployment.
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-destinations/destination-syslog-ng-otlp/_index.md

@@ -72,6 +72,8 @@ The `syslog-ng-otlp()` destination has the following options.
 
 *Description:* The URL of the {{% param "product.abbrev" %}} receiver.
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-destinations/google-bigquery/_index.md

@@ -193,6 +193,8 @@ Alternatively, you can set the schema with the [`protobuf-schema()`](#protobuf-s
 
 *Description:* The URL of the Google BigQuery where the logs are sent.
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 
 {{< include-headless "chunk/option-destination-threaded-workers.md" >}}

content/chapter-destinations/google-pubsub-grpc/_index.md

@@ -176,6 +176,8 @@ auth(
 
 *Description:* An alias for [`service_endpoint`](#service-endpoint).
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-destinations/opentelemetry/_index.md

@@ -111,6 +111,8 @@ log non_otel_to_otel_tls {
 
 *Description:* The URL of the OpenTelemetry receiver.
 
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+
 <a id="worker-partition-key"></a>
 {{< include-headless "chunk/option-destination-http-worker-partition-key.md" >}}
 

content/chapter-nonsequential-processing/_index.md

@@ -57,3 +57,7 @@ log {
 ```
 
 Staring with {{< product >}} version 4.17, you can use the `batch-size()` option to specify how many consecutive messages should be processed by a single `parallelize()` worker. This ensures that this many messages preserve their order on the destination side, and also improves `parallelize()` performance. A value around 100 is recommended for `batch-size()`. Default value: `0` (batching is disabled).
+
+<!-- FIXME mention/explain
+{{< include-headless "chunk/option-destination-worker-partition-buckets.md" >}}
+ -->

content/headless/chunk/option-destination-worker-partition-buckets.md

@@ -0,0 +1,11 @@
+---
+---
+<!-- This file is under the copyright of Axoflow, and licensed under Apache License 2.0, except for using the Axoflow and AxoSyslog trademarks. -->
+## worker-partition-buckets()
+
+|          |        |
+| -------- | ------ |
+| Type:    | template |
+| Default: |        |
+
+*Description:* The `worker-partition-buckets()` option determines the number of worker threads used for the `worker-partition-key()`. Note that the number set by `worker-partition-buckets()` should be lower than the number of `workers()`.

content/whats-new/_index.md

@@ -15,6 +15,8 @@ weight: 10
     - You can now set a suffix for your S3 objects using the [`object-key-suffix()` option]({{< relref "/chapter-destinations/destination-s3/_index.md#object-key-suffix" >}}).
     - You can change the default checksum settings for S3 compatible solutions that don't support checksums using the [`use-checksum()` option]({{< relref "/chapter-destinations/destination-s3/_index.md#use-checksum" >}}).
 
+- The `http()` and other threaded destinations now have a [`worker-partition-buckets()` option]({{< relref "/chapter-destinations/configuring-destinations-http-nonjava/reference-destination-http-nonjava/_index.md#worker-partition-buckets" >}}) that determines the number of worker threads used for the `worker-partition-key()`.
+
 ## Version 4.19 (2025-10-15)
 
 - The [`dict_to_pairs`]({{< relref "/filterx/function-reference.md#dict-to-pairs" >}}) FilterX function can convert a dictionary to a list of pairs.