You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Multiple scheduled daily workflows fire at the same minute (e.g., 12:05 UTC), causing a burst of concurrent GitHub installation token API calls that exhausts the 60 req/minute quota. Affected workflows' safe-outputs step silently loses work: issues that were fully prepared by the agent are never written to GitHub. On 2026-05-01, this caused at least 4 workflows to lose their safe-output writes during a 12:05–12:28 UTC window. A similar incident occurred on 2026-04-30 (#29318).
Recommended fix: Add ±0–5 minute jitter to cron expressions for all scheduled daily workflows so that the burst of concurrent starts is spread over a wider window. The existing retry logic (3 attempts over ~90s) is insufficient when the token is completely drained.
Eliminates silent safe-output loss in concurrent burst windows, improving data integrity and reducing the ~10% re-occurrence rate of rate-limit bursts.
Suggested Agent
Existing workflow refactoring agent, or a new agent that can systematically update cron expressions across all scheduled workflows.
Estimated Effort
Medium (1–4 hours) — enumerate all daily-schedule workflows, stagger their cron times by random offsets.
Activity
github-actions commentedon May 1, 2026
🍪 Issue Monster selected this for Copilot
I've identified this issue as a good candidate for automated resolution and requested assignment to the Copilot coding agent.
If assignment succeeds, the Copilot coding agent will analyze the issue and create a pull request with the fix.
Om nom nom! 🍪