Repository navigation
[agent-job-health] Agent Job Health Report, 2026-11-13 (partial coverage) #77583
Closed
Replies: 1 comment
|
This discussion has been marked as outdated by Agent Job Health Monitor. A newer discussion is available at Discussion #77876. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Summary
Note
Coverage is partial: only about 7 of the 24 hours were collected. Treat these figures as a subset, not the full day.
Failure Rate by Step
Tracked Failures
None confirmed. Issue search results for the three workflows were hidden by the integrity filter (10 hidden in each of two searches, 9 in the third). The one visible match, open issue #72944 (PR Sous Chef add-comment fails with Bad credentials), describes a safe-output credential problem rather than a Pi CLI execution failure, so it is not counted as tracking this failure.
Novel Failure Clusters
No cluster meets the issue threshold (two or more workflows, or 10% or more of runs). The three failures share no signature: each is one run in a different workflow, engine, and step. Candidates pending classification:
Error excerpts could not be retrieved. Job logs were blocked by the integrity policy, and the local audit files give only error counts. Blast radius: single workflow each, with no engine-wide or fleet-wide signal in this subset.
Schedule Heartbeat
Partial check. evoskill-evolver is cron
20 3 */21 * *(source says tri-weekly). Last run 2026-11-01T03:26Z, success. Gap 12.8 days against a threshold of 43 days (2x21 plus 1). Not a blind spot. Next run due about 2026-11-22.Not checked this run: the other schedule-triggered workflows (about 250 in the 2026-11-10 population). Each listing call returns about 3 KB, and the local logs cover only about 7 hours, so logs cannot show that a workflow is silent. The 2026-11-10 sweep found 0 blind spots. Blind spots in the checked subset: none. Full sweep outstanding.
View per-workflow breakdown
Mapped run counts are incomplete. Workflows with a failure:
Recommendations
Issue Creation
No issue created. The novel-cluster threshold is not met, and the checked heartbeat subset found no blind spot.
Method and Limits
All reactions