[tatoalo]

Problem

cloud task streams can churn when a live stream emits keepalives and then disconnects, or when the client loses access to task status/log endpoints.
keepalives were resetting the reconnect attempt counter, and 401/403/404 status or log responses were treated like generic transient failures as showni in slack thread logs

Changes

• move the reconnect-attempt reset so only non-keepalive SSE events mark the stream as healthy
• fail the cloud task watcher on 401, 403, or 404 responses from status and persisted-log fetches, using the same user-facing stream error details as the SSE path.

[greptile-apps]

Prompt To Fix All With AI

Fix the following 1 code review issue. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 1
apps/code/src/main/services/cloud-task/service.test.ts:602-668
**Parameterise over all three guarded status codes**

`shouldFailWatcherForFetchStatus` guards against 401, 403, and 404, but both new tests only exercise 403. 401 (expired token) and 404 (run deleted) are distinct failure modes that are easy to miss if the function is ever changed.

The two tests also share exactly the same assertion shape and could be collapsed into an `it.each` table, matching the project's preference for parameterised tests.

_{Reviews (1): Last reviewed commit: "fix(cloud-agent): handle stream churn" | Re-trigger Greptile}