fix: preserve error traces

[evan-onyx]

Description

Some stack traces were being overwritten with unhelpful errors, this change allows them to be preserved

How Has This Been Tested?

n/a, unlikely this can cause problems

Backporting (check the box to trigger backport action)

Note: You have to check that the action passes, otherwise resolve the conflicts manually and tag the patches.

• This PR should be backported (make sure to check that the backport attempt succeeds)
• [Optional] Override Linear Check

---

Summary by cubic

Fixed an issue where error stack traces were being overwritten, so original error details are now preserved during task failures.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
internal-search	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jul 25, 2025 5:49pm

[greptile-apps]

Greptile Summary

This PR fixes a bug where valuable error information was being lost during the index attempt failure handling process. The change adds a safety check in the docfetching task watchdog to prevent overwriting existing error traces when marking failed index attempts.

The core issue was that when a spawned indexing task failed, the watchdog would unconditionally try to mark the attempt as failed with a generic "spawned task exceptioned" message, even if the actual indexing process had already marked itself as failed with detailed error information. This resulted in the loss of real stack traces and error details that would be crucial for debugging.

The fix introduces a terminal state check using attempt.status.is_terminal() before attempting to mark the failure. This ensures that once an index attempt reaches a terminal state (SUCCESS, FAILED, CANCELED, COMPLETED_WITH_ERRORS) with its original error information intact, that information is preserved rather than being overwritten.

This change integrates well with the existing error handling infrastructure in the codebase, particularly the IndexAttempt model's status tracking and the mark_attempt_failed function that handles failure recording with stack traces.

Confidence score: 5/5

• This change is extremely safe to merge and will not cause any production issues.
• The change only adds a protective guard condition that prevents data loss without altering core functionality.
• No files require additional attention - the change is isolated and straightforward.

_{1 file reviewed, no comments}

_{Edit Code Review Bot Settings | Greptile}

[cubic-dev-ai]

cubic analysis

No issues found across 1 file. Review in cubic