PrefectHQ
prefect
Blog
Docs
Changelog
Blog
Docs
Changelog
Overview
Optimizations
Branches
Benchmarks
Runs
Performance History
Latest Results
chore: replace Sanity logo URLs Co-authored-by: Alex S. <alex.s@prefect.io> Co-Authored-By: alex.s <ajstreed1@gmail.com>
devin/1790715382-replace-logo-urls
2 hours ago
assets: add repository-hosted block and worker logos (#23231) Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com> Co-authored-by: Alex S. <alex.s@prefect.io> Co-authored-by: alex.s <ajstreed1@gmail.com>
main
2 hours ago
Fix pyright errors in server_maintenance (OSS-8278)
jimi/worker-healthy-during-maintenance
3 hours ago
Fix import ordering. Signed-off-by: Rich Scott <richscott@sent.com>
richscott:richscott/armada-integration
3 hours ago
Count the worker health window from the end of the last maintenance back-off The previous approach reported healthy only while a maintenance back-off was in progress (plus the request timeout). After the lift, the last poll was still as old as the whole window, so a plain 502/503 or timeout on the first post-lift retry flipped /health to 503 straight away, and a liveness probe could restart the worker while it recovered. Record when each maintenance sleep ends, keeping the latest, and measure staleness from the later of the last poll and that time against the usual polling threshold. This gives the full window of grace after the last back-off and removes the special case. The process-wide record moves to the private module prefect._internal.server_maintenance so the Runner can read it later without importing the client. The duplicated maintenance check in the async and sync retry loops is folded into one helper. Tests drive the real client and worker with respx maintenance responses instead of setting module state. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
jimi/worker-healthy-during-maintenance
4 hours ago
Merge branch 'main' into richscott/armada-integration
richscott:richscott/armada-integration
4 hours ago
Fix `get_call_parameters` to unwrap `FieldInfo` defaults for flow runs (#23228)
main
5 hours ago
Install locked dependency versions in the Docker image (#23225)
main
6 hours ago
Latest Branches
CodSpeed Performance Gauge
0%
chore: replace Sanity-hosted logo URLs
#23235
2 hours ago
56c387e
devin/1790715382-replace-logo-urls
CodSpeed Performance Gauge
0%
Keep the worker healthcheck healthy while the API is in maintenance
#23230
3 hours ago
02c8932
jimi/worker-healthy-during-maintenance
CodSpeed Performance Gauge
0%
Add work pool support for Armada (K8s multi-cluster) batch system
#23232
4 hours ago
6321dfb
richscott:richscott/armada-integration
© 2026 CodSpeed Technology
Home
Terms
Privacy
Docs