BulkPartialIndexWorker — partial update before full document indexed

Now I'll invoke the RCA skill to analyze this cluster.

Let me search Datadog for more context about this error and explore the source code in parallel.

The error message "ES doc_missing after max retries" comes from a `DocMissingRetryWorker` but wasn't

Not in the current checkout. Let me check for recent deployments or branches, and also search in the

Let me check if this worker was recently added and might be on the deployed branch.

The worker isn't in the develop branch either. The deploy version was `production-us-west-2-20260519

Found it! The file exists at `app/workers/doc_missing_retry_worker.rb` on the latest develop. Let me

Excellent! Now I have the full worker code. Let me find what calls this worker — specifically the up

Now I need to find what calls `BulkPartialIndexWorker` — specifically the code that partial-updates

Now let me understand the flow better. The issue is that `ElementTrace` records are updated in the d

Now let me understand the key part - the initial indexing. Let me look at how ElementTrace gets inde

Now I have a clear picture. Let me also verify the TSLA-12834 commit that introduced the partial upd

Now I have a comprehensive understanding. Let me search Datadog for one more piece of context - whet

Now I have a comprehensive picture. Let me investigate one more piece — the `after_commit` callback

Now let me look at the key piece — how ElementTrace records are created in the context of `CreateSit

Now let me look at `assign_editing_to_editing_entity` and `_stamp_editing_id_on_elements` to underst

Now I have a complete understanding. Let me check one more thing — how ElementTrace gets created and

This is the crucial piece. I can see the flow now:

I now have enough evidence to write the RCA. The root cause is clear: there's a **race condition** b

Now the report is written. Per the instructions, I should NOT modify the cluster frontmatter (the us

RCA report written to `./content/docs/incidents/e671a756-22ee-4ef8-a0ec-55aeb1106087/rca.mdx`.

Now I have the cluster info. Let me invoke the RCA skill to perform the analysis.

Now let me search Datadog logs and explore the source code. The cluster is for `cupixworks-worker` s

Let me check if there's an existing RCA for this same error pattern, and also look at the specific c

These workers aren't in the current codebase — they might be a newer deployment or in a different lo

The code was added recently. Let me check the latest commit that added this feature.

Now let me check the upstream code that calls `BulkPartialIndexWorker` — the code that queues partia

Now let me find what calls `BulkPartialIndexWorker` — the PR that introduced the inline bulk call.

The `BulkPartialIndexWorker` files aren't on the current local branch — they're in a newer commit. L

Now let me look at where `BulkPartialIndexWorker` is called — the callers that trigger the partial u

Now I have the full picture. Let me also check if there's an existing RCA for a similar issue.

The directory already exists with a previous RCA. Let me check the previous RCA to see if it covers

This is a companion cluster with the same root cause. Now I have all the evidence. Let me write the

The file exists but has old/incomplete content. Let me write the full RCA report now.

Now let me verify the cluster file frontmatter to confirm the current `rca_status` — but per instruc

RCA report written to `./content/docs/incidents/e671a756-22ee-4ef8-a0ec-55aeb1106087/rca.mdx`.