PrefectHQ/prefect - CodSpeed

prefect

Blog Docs Changelog

Performance History

Latest Results

Deduplicate Kubernetes observer state checks and diagnosis logs for unchanged Pending pods (#22965) Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: Alexander Streed <alex.s@prefect.io>

main

10 hours ago

Use indexed flow run cleanup with separate orphan reconciliation (#22875) Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: Alexander Streed <alex.s@prefect.io>

main

12 hours ago

Keep orphan vacuum cleanup automatic

devin1/oss-8158-db_vacuum-deletes-flow-runs-first-then-full-scans-log-anti

14 hours ago

Docs: Disable logging (#23017)

main

14 hours ago

Crash flow runs interrupted by worker shutdown `_submit_run_and_capture_errors` handled only `Exception`. Shutdown cancels that frame and the cancellation exception derives from `BaseException`, so it passed over the handler and no terminal state was ever proposed: the run stayed `Running` for ever and kept its work pool concurrency slot on the server, while the worker's own `finally` released the local one and left it looking healthy. Propose `Crashed` from a cancellation handler before re-raising, only for a run that had actually started, and shield the call because the scope is cancelled by the time it runs. A run that is already terminal is unaffected: `propose_state` raises `Abort` and `_propose_crashed_state` ignores it. Reaching a terminal state is the scope here. Making the worker drain rather than cancel is a separate change and is what #22528 asks for.

alekseevpavel04:fix/crash-runs-on-worker-shutdown

16 hours ago

Retarget the URL health-check test at the redis_from_url seam `get_async_redis_client` now builds every URL-backed client through `prefect_redis.connection.redis_from_url`, so the falsy-override test added in #22974 no longer observes the patched `prefect_redis.client.Redis`. Patch the factory the URL branch actually calls; the assertion that an explicit `health_check_interval=0` survives the settings fallback is unchanged, and `redis_from_url`'s own kwarg pass-through is covered by `tests/test_connection.py`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

fatih-acar:feat/redis-sentinel-on-22211

16 hours ago

Fix flow run vacuum cleanup races

devin1/oss-8158-db_vacuum-deletes-flow-runs-first-then-full-scans-log-anti

16 hours ago

Bump browserslist from 4.25.1 to 4.28.8 in /ui-v2 (#23010) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

main

18 hours ago

Latest Branches

0%

Use indexed flow run cleanup with separate orphan reconciliation#22875

14 hours ago

7bcdf91

devin1/oss-8158-db_vacuum-deletes-flow-runs-first-then-full-scans-log-anti

0%

Crash flow runs interrupted by worker shutdown#23016

17 hours ago

04388b2

alekseevpavel04:fix/crash-runs-on-worker-shutdown

0%

Add Redis Sentinel support to prefect-redis#22377

16 hours ago

60142d7

fatih-acar:feat/redis-sentinel-on-22211

© 2026 CodSpeed Technology

Home Terms Privacy Docs