Timeout-then-queue background search

[level09]

Summary

Long searches no longer hold a web worker (and its GIL) hostage. Interactive searches on the bulletin, actor, and incident list endpoints now run under a Postgres statement_timeout (default 30s, SEARCH_TIMEOUT env, 0 disables). When the database cancels one, the same query is re-queued to Celery, which re-runs it with a generous bound (default 10 min, BACKGROUND_SEARCH_TIME_LIMIT), stores up to 10k matching ids in Redis for 24 hours, and notifies the user with a link that replays the stored ids in the normal list view via the existing ids filter.

This is the same sync-with-async-fallback model used by Elasticsearch's async search API (wait_for_completion_timeout) and Superset's SQL Lab (SQLLAB_TIMEOUT + Celery).

• enferno/utils/background_search.py: timeout helper (SET LOCAL, so the bound dies with the transaction and can't leak through the pool), 57014 detection, Redis result store, and a timeout_fallback decorator that converts a canceled search into a 202 {queued, token} response
• enferno/tasks/background_search.py: Celery task; id-only select, completion/failure notification (new BACKGROUND_SEARCH_STATUS event, in-app only)
• The three list endpoints get the decorator plus one apply_search_timeout() call placed after activity logging (the activity commit would otherwise end the transaction and drop the bound)
• GET /admin/api/background-search/<token>: owner-only stored ids
• Frontend: 202 shows a snackbar; ?bgs=<token> on the list pages loads stored ids and runs the search

Test plan

• uv run pytest tests/test_background_search.py (timeout detection via real pg_sleep cancel, result store round-trip, full task run asserting stored ids + notification, endpoint 202 path)
• tests/test_core_entities.py and tests/test_search_query_combination.py pass unchanged
• Manual: set SEARCH_TIMEOUT=1, run a heavy search, confirm snackbar, notification, and results link

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 8f1778ff-1ab9-4062-b485-9efd1cf2e264

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/background-search

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}