v1.6: Limit task batch size (#2667)

config/sidebar-learn.json

@@ -408,6 +408,11 @@
  "label": "Overview",
  "slug": "overview"
  },
+ {
+ "source": "learn/experimental/max_number_of_batched_tasks.mdx",
+ "label": "Limit number of batched tasks",
+ "slug": "max_number_of_batched_tasks"
+ },
  {
  "source": "learn/experimental/metrics.mdx",
  "label": "Metrics",

learn/experimental/max_number_of_batched_tasks.mdx

@@ -0,0 +1,26 @@
+---
+title: Limit task batch size — Meilisearch documentation
+description: Use this experimental feature to limit the number of tasks Meilisearch performs in a single batch. May improve system stability in some cases.
+---
+
+# Limit task batch size
+
+Use this experimental feature to limit the number of tasks Meilisearch performs in a single batch. May improve stability in systems handling a large queue of resource-intensive tasks.
+
+To enable it, relaunch Meilisearch with the `--experimental-max-number-of-batched-tasks` command-line option, specifying the maximum amount of tasks you want to process in a single batch:
+
+```sh
+./meilisearch --experimental-max-number-of-batched-tasks 100
+```
+
+You may also activate this option with environment variables or in the configuration file by using `MEILI_EXPERIMENTAL_MAX_NUMBER_OF_BATCHED_TASKS`.
+
+[Visit the GitHub discussion to learn how to use this feature](https://github.com/orgs/meilisearch/discussions/713).
+
+Are you using this feature? Meilisearch wants to hear from you! Tell us about your experience in the [GitHub discussion](https://github.com/orgs/meilisearch/discussions/713). All feedback is useful and helps make Meilisearch better and easier to use.
+
+<Capsule intent="warning">
+This is an experimental feature. Experimental features are unstable: their API might significantly change and become incompatible between releases. Meilisearch does not recommend using experimental features in a production environment.
+
+Meilisearch makes experimental features available expecting they will become stable in a future release. However, it is not possible to guarantee when and if this will happen.
+</Capsule>

learn/experimental/overview.mdx

@@ -44,8 +44,9 @@ Activating or deactivating experimental features this way does not require you t
 
 | Name | Description | How to configure |
 |----------------------------------------------------------------------------------|------------------------------------------------------------|---------------------------------------------------|
+| [Limit task batch size](/learn/experimental/max_number_of_batched_tasks) | Limits number of tasks processed in a single batch | At launch with a CLI flag or environment variable |
 | [Metrics API](/learn/experimental/metrics) | Exposes Prometheus-compatible analytics data | At launch with a CLI flag or environment variable, during runtime with the API route |
+| [Puffin reports](/learn/experimental/puffin_reports) | Export profiler data to diagnose performance issues | During runtime with the API route |
 | [Ranking score details](/learn/experimental/ranking_rule_score_details) | Exposes advanced result ranking score data | During runtime with the API route |
 | [Reduce indexing memory usage](/learn/experimental/reduce-indexing-memory-usage) | Optimizes indexing performance | At launch with a CLI flag or environment variable |
 | [Vector search](/learn/experimental/vector_search) | Allows Meilisearch to function as a vector embedding store | During runtime with the API route |
-| [Puffin reports](/learn/experimental/puffin_reports) | Export profiler data to diagnose performance issues | During runtime with the API route |