macos-latest ARM64 jobs cancelled with zero steps after hosted-runner acquisition failures #211885
Replies: 18 comments 4 replies
|
💬 Your Product Feedback Has Been Submitted 🎉 Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users. Here's what you can expect moving forward ⏩
Where to look to see what's shipping 👀
What you can do in the meantime 💻
As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities. Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐ |
|
Same issue. |
|
Happens also on public repos. https://github.com/tbeu/matio/actions/runs/41673532046/job/126047923070
Still, githubstatus.com shows no incidents for GH Actions which obviously is fake data. |
|
same here |
|
Workaround that worked for me: pinning the runner to macos-15 instead of macos-latest (which currently maps to the macos-26 arm64 image). The job got a runner instntly, while macos-latest failed 3 times in a row. Maybe works for you aswell... |
|
I also keep running into this issue. Seems impossible to run a macOS arm64 runner in the evening (CET) now. I gave up after retry attempt # 9. In the morning it works fine, but it does mean I have the schedule my workflow to the mornings |
|
I have the same issue in https://github.com/MarketSquare/robotframework-browser/actions/runs/41904244509/job/126834528053 Have been same problem at least from yesterday. Problem seems to come and go.
|
|
Same issue right here, really difficult to release properly since 3 days... |
|
I'm seeing the same for a few days now |
|
Additional confirmation of the issue — not a proposed solution. We are experiencing the same issue with GitHub-hosted macOS Arm64 runners on November 5, 2026. Across five consecutive workflow runs, our Linux JVM verification jobs completed successfully, while the macOS iOS simulator jobs were cancelled before executing any steps. Our workflow annotations report: "The job was not acquired by Runner of type hosted even after multiple attempts" We also received an internal server error and repeated macOS Arm64 capacity warnings. Could GitHub confirm whether this is an ongoing hosted-runner capacity or scheduling issue, and whether a reliable workaround or estimated resolution is available? This is currently blocking our Kotlin Multiplatform iOS verification. |
|
For smaller CI (e.g. check only with the latest available Python interpreter) one can manually launch (again) such a failed run. But this approach does not scale well, let alone is not the intended automation (especially if the results are input for a subsequent workflow). |
|
Same here today. |
|
Still having this issues after multiple attempts. |
|
We are seeing the same failure in Knowledge Retrieval Engine on November 16, 2026 with both GitHub-hosted ARM64 labels. The affected jobs were cancelled before receiving a runner or executing any test steps: GitHub Actions reported that the jobs were not acquired by a hosted runner after multiple attempts and warned that macOS ARM64 jobs may queue longer due to capacity constraints. The run also reported an internal server error. The other CI jobs that executed passed, including the public package installation checks, and CodeQL passed separately. Can GitHub confirm whether there is an ongoing hosted macOS ARM64 runner-capacity or allocation issue and recommend a reliable supported resolution that preserves ARM64 coverage? |
|
Still happening to me today. And when I tried reaching out to github support, they closed the ticket, and told me my account doesn't suppor this type of support. |




Uh oh!
There was an error while loading. Please reload this page.
🏷️ Discussion Type
Bug
💬 Feature/Topic Area
Actions Runner
Discussion Details
On 2 November 2026 we repeatedly observed standard macos-latest ARM64 jobs remain queued for about 15 minutes and then be cancelled before any step ran. Job metadata reported runner_id: 0 and an empty steps array. The workflow conclusion was failure, while the job conclusion was cancelled.
Annotations reported:
Examples occurred around 16:21–16:36, 18:00–18:15 and 19:00–19:15 UTC. Other jobs did receive a macos-26-arm64/20261026.0455 runner between these failures, so allocation appears intermittent. We have stopped repeated unchanged retries. No checkout or application test ran in the cancelled jobs.
The public status page reports all systems operational. Is there a known incident affecting this specific standard ARM64 pool, and what is the recommended recovery/retry guidance while keeping the same original runner environment? Please also clarify whether any account-level scheduling condition could produce these exact annotations.
The repository is private. This public report intentionally excludes repository names, run links/IDs, source code, credentials and personal project data. We can provide scheduling identifiers through an appropriate private support channel if required.
Thank you.
All reactions