CapturesController#index Redis cache miss causing excessive DB joins

Now I have the cluster details. Let me invoke the RCA skill.

I'll analyze this latency cluster for `Api::V1::CapturesController#index`. Let me gather context and

Let me search Datadog logs and explore the source code in parallel.

Excellent. I have comprehensive data from both searches. Now let me look at the `AccessibleEntities:

Now let me look at how the `_search` method works in the `BaseRepository` and how it integrates with

Now let me check how `AccessibleEntities::Cache` is invoked in the request flow — it's being called

Now let me find where these cached permission methods are called during the CapturesController#index

Let me find where the `AccessibleEntities::Cache` flush is triggered during the request — likely in

Now I need to understand the Elasticsearch query that's built for captures, specifically how it uses

Now let me find the Elasticsearch query builder for captures to understand how it uses permission ID

The `AccessibleEntities::Cache` flushing happens at the User model level as a cache warming mechanis

Now let me find the `_search` method in CaptureRepository to understand the Elasticsearch query buil

This is the key finding. At line 669, the Elasticsearch query includes:

Now let me check the `readable_record_ids` method in the `AccessibleEntities` module (line 669 calls

Now let me check the `DEFAULT_PERMISSION_CACHE_EXPIRES_IN` constant and `directly_permitted_items` t

Excellent! Now I have all the evidence I need. The key findings are:

Now let me update the cluster frontmatter. The user said NOT to modify the frontmatter — the pipelin

RCA complete. Report written to `./content/docs/incidents/6ea5c685-591f-4611-8009-08c0bbdc0b80/rca.m