PrefectHQ/prefect - CodSpeed

prefect

Blog Docs Changelog

Performance History

Latest Results

Remove inline comments Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

fix-workspace-resolver-pull-step-logs

35 minutes ago

Trim comment to the load-bearing invariant Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

fix-workspace-resolver-pull-step-logs

52 minutes ago

Fix cleanup registry tests: add work_pool_updated to make_cleanup_connection Co-authored-by: Alex Streed <alex.s@prefect.io> Co-Authored-By: alex.s <ajstreed1@gmail.com>

devin/1782998996-fix-stale-work-pool-template

2 hours ago

Fix asyncio.TimeoutError compatibility on Python 3.10 On Python 3.10, asyncio.TimeoutError is a separate class from the builtin TimeoutError (unified in 3.11+). asyncio.wait_for raises asyncio.TimeoutError, so the except clause must catch that explicitly. Also moves get_current_settings import to module level per conventions. Co-authored-by: Alex Streed <alex.s@prefect.io> Co-Authored-By: alex.s <ajstreed1@gmail.com>

devin/1782998996-fix-stale-work-pool-template

3 hours ago

Add server-side periodic snapshot reconciliation on idle worker channels When no snapshot invalidation arrives within the configured reconciliation window (default 300s), the server re-reads the work pool and pushes a fresh snapshot over the channel. This ensures workers self-heal from missed invalidation events without reintroducing REST polling on the worker side. The _send_loop now uses asyncio.wait_for with a timeout on the snapshot queue. On timeout, it synthesizes a 'periodic_reconciliation' invalidation that flows through the existing snapshot-build-and-send path. The interval is configurable via PREFECT_SERVER_WORKER_CHANNEL_SNAPSHOT_RECONCILIATION_SECONDS. Closes #21885 Co-authored-by: Alex Streed <alex.s@prefect.io> Co-Authored-By: alex.s <ajstreed1@gmail.com>

devin/1782998996-fix-stale-work-pool-template

3 hours ago

Move REST reconciliation timer update to after successful sync Update _last_rest_sync_monotonic only after the REST sync completes successfully (or is superseded by a channel snapshot). This ensures transient REST failures trigger an immediate retry on the next heartbeat rather than waiting another full reconciliation interval. Co-authored-by: Alex Streed <alex.s@prefect.io> Co-Authored-By: alex.s <ajstreed1@gmail.com>

devin/1782998996-fix-stale-work-pool-template

4 hours ago

chore(deps-dev): bump the eslint group in /ui-v2 with 2 updates (#22423) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

main

4 hours ago

chore(deps): bump the uv-dependencies group with 6 updates (#22427) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

main

4 hours ago

Latest Branches

+1%

Send workspace resolver pull step logs to the API#22432

50 minutes ago

776b2f2

fix-workspace-resolver-pull-step-logs

0%

Reconcile stale work-pool snapshots on worker heartbeat#22429

2 hours ago

40f7209

devin/1782998996-fix-stale-work-pool-template

0%

Preserve a paused deployment schedule on redeploy#22396

3 days ago

18b4d90

apoorva-01:fix/19302-preserve-disabled-schedule-on-redeploy

© 2026 CodSpeed Technology

Home Terms Privacy Docs