GitHub Pages run remains Queued after cancellation request (unionmed/sayd-magazine)

[unionmed]

🏷️ Discussion Type

Question

💬 Feature/Topic Area

Workflow Deployment

Discussion Details

My GitHub Pages deployment is stuck in Queued and is blocking publication.

Public run: https://github.com/unionmed/sayd-magazine/actions/runs/42594893923
Repository: unionmed/sayd-magazine
Workflow: pages-build-deployment, run #522, latest attempt #3
Commit: 1aca6d2c56ee39aa2901cbe6c9d54a71c8511137 (main)

The current attempt was triggered on November 9, 2026 at 10:04 UTC and still shows Queued on November 10. Earlier attempts #1 and #2 showed failed. A check of the latest attempt through the Actions API returned an empty jobs array.

On November 9, I used Cancel run in the signed-in Actions UI. The menu closed, but cancellation was not confirmed and the run remains Queued. I have not tried force-cancel and do not have an HTTP error code to report. I have not started another deployment while this run remains unresolved.

What is the safe next step to clear this stuck attempt and rerun the same commit without creating competing Pages deployments? Can GitHub staff inspect the run, or point me to the appropriate escalation route?

[zero7goat]

This looks more like a stuck Actions/Pages queue issue than a problem with the commit itself.

I'd try cancelling the run through the Actions REST API first. If the normal cancellation doesn't work, the force-cancel endpoint may be worth trying:

gh api -X POST \
  repos/unionmed/sayd-magazine/actions/runs/42594893923/cancel

If that doesn't clear it, try:

gh api -X POST \
  repos/unionmed/sayd-magazine/actions/runs/42594893923/force-cancel

Since the API is returning an empty jobs array, it would also be useful to check whether the run is stuck before job creation rather than waiting for a runner.

If both cancellation methods fail, I'd avoid triggering another Pages deployment for now. Open a GitHub Support ticket with the run URL, repository, commit SHA, timestamps, and cancellation responses. This may require GitHub to inspect the backend queue.

Also, check [GitHub Status](https://www.githubstatus.com/) for any ongoing Actions or Pages incidents.

The important part is to get the existing run cleared before trying to publish the same commit again.

[punit2911]

UI cancel didn’t work, so use force-cancel through the REST API, confirm the run is finished, then re-run the same run instead of pushing a new commit.

1. Check for an incident first. Open githubstatus.com and look for an Actions or Pages incident. If there is one, the queue may clear by itself.

2. Force-cancel the stuck run. Force-cancel is meant for runs that ignore a normal cancel, and it needs write access to the repo. You can run this with the GitHub CLI:

gh api --method POST repos/unionmed/sayd-magazine/actions/runs/42594893923/force-cancel

3. Confirm it actually ended:

gh api repos/unionmed/sayd-magazine/actions/runs/42594893923 --jq '.status, .conclusion'

You want completed and cancelled. Don’t re-run before this.

4. Re-run the same run (the same commit, 1aca6d2) from the Actions UI with Re-run all jobs, or with:

gh run rerun 42594893923 --repo unionmed/sayd-magazine

A re-run makes a new attempt of the same run. A new push or empty commit would start a second pages-build-deployment that queues behind the stuck one, which is the competing deployment you want to avoid.

Why the UI cancel did nothing: your latest attempt has an empty jobs array, so it never got jobs assigned. A normal cancel has nothing to act on, which points to a platform-side stuck state and not a workflow problem.

5. If force-cancel fails, escalate to GitHub Support. Others with the same problem reported force-cancel returning HTTP 500 or 409, and they had to go to Support. Community discussions can’t get staff to inspect an individual run, so open a ticket at support.github.com/contact. Include:

the repo and the run URL and ID
the exact commands and full HTTP responses
the time it got stuck, with timezone
that the repo is public