Api::V1::SketchesController#show (avg 10131ms, max 10131ms)

Now I have the cluster details. This is a latency cluster for `Api::V1::SketchesController#show` tak

Let me search Datadog logs and explore the source code for this latency issue.

Interesting! I can see Elasticsearch timeout errors around the same time (14:21 KST, which is 05:21

I can see the pattern: at 14:21 KST (05:21 UTC), there were multiple Elasticsearch timeout errors (1

The `show` action is handled via `set_sketch` before_action (line 10) which calls `repository_instan

Now let me look at the `BaseRepository#show` method and the `Invokable::SketchesController` module t

Now let me look at the class-level `show` method in `BaseRepository` to understand how it queries da

Now I need to understand what happens in the `show` action of the controller — looking at what `ApiC

So the `show` action simply renders `@model` which is set by `set_sketch` before_action. The latency

Now let me check the `Sketch` model for any Elasticsearch callbacks:

The `Sketch` model includes `EntityIndexable`. This is likely the Elasticsearch integration. Let me

Now let me also check the `Searchable::Sketch` module which might handle Elasticsearch searches:

Now let me check the `Searchable` module to understand the `after_commit` callbacks:

Now I have a clear picture. The `Searchable` module's `_update_document` method (line 55) is trigger

This is a key finding. `set_current_workspace` at line 58 calls `update(current_workspace_id: worksp

There it is. Line 25: `timeout: 10` — this is a 10-second timeout for Elasticsearch requests. This c

Confirmed. So the flow is:

All the errors cluster at 14:21 KST (05:21 UTC) — exactly when our trace was captured. This confirms

Now let me create the RCA report:

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

RCA report written to `./content/docs/incidents/4becc8fc-599b-421f-b333-e31366f32918/rca.mdx`.