Stuck "Summary" check runs permanently in_progress — no way to cancel or recover

[altendky]

I have Mergify "Summary" check runs stuck in in_progress across multiple PRs:

• PR #283 check run 67443866010 — stuck since 2026-03-17T17:07Z
• PR #280 check run 67457336195 — stuck since 2026-03-17T18:35Z

Both show status: in_progress, conclusion: null, output title "Your rules are under evaluation", summary "Be patient, the page will be updated soon." They never complete.

I have a 90-minute merge queue check timeout configured, but I have not observed queued or dequeued labels being applied to these PRs during this exploration — it seems like they never entered the queue in the first place.

What I have tried (none had any effect):

• Removing and reapplying my enqueue label — the stuck check run was not updated.
• @mergifyio refresh — no apparent effect on the stuck check run.
• GitHub API (PATCH /check-runs/{id} with status: completed, conclusion: cancelled) — returns 403: You must authenticate via a GitHub App. Only the Mergify app can update its own check runs, so repo owners cannot cancel them.
• Empty commit + push to create a new check run on a fresh SHA — this did create a new check run (67470059342), but that one is also stuck in the same in_progress state. PR Add file-reference mechanism for tool parameters altendky/onshape-mcp#280 now has two stuck Summary check runs simultaneously — the original from the real commit and the new one from the empty commit — both permanently in_progress.

Questions:

1. Is there any way for a repo owner to cancel or force-complete a stuck Mergify check run? None of the approaches I tried worked. The GitHub API restriction means only the Mergify app can update these, leaving users with no escape hatch.

2. Is there a timeout for these check runs? Will they eventually self-resolve, or will they stay in_progress indefinitely?

3. Is there somewhere else I can get more information about the status of these runs? The check run output just says "Be patient" with no additional details. The Mergify dashboard event logs do not seem to show anything actionable.

4. Is there a workaround I am missing? The empty commit trick did not help since new check runs also get stuck, suggesting something is broken in Mergify's processing of my repo specifically.

Thanks for your time and help.

Cheers,
-kyle

[JulianMaurin]

Hi Kyle,

Thank you for the detailed report.

Yesterday (March 17) we experienced an infrastructure incident between approximately 16:37 and 21:30 UTC that affected repositories on our open-source plan. During this window, Mergify was unable to process events or update check runs for affected repositories, which is exactly what you observed. Summary check runs stuck as in_progress with no way to move forward.

Once the incident was resolved, your PRs were processed normally: PR #280 entered the queue at 21:40 UTC and merged at 21:51 UTC, and PR #283 entered at 21:51 UTC and merged at 22:09 UTC.

We apologize for the disruption and are taking steps to prevent this class of failure from recurring.

[altendky]

thanks for the explanation!