Valid labels for Github-hosted runners on architecture ARM/ARM64

[brainchild0]

I am working with a codebase that has been deployed successfully on MacOS and Linux.

I wish to expand the range of targets for the automatic testing provided by the workflow for Github.

I expected to be successful inserting the following in the specification for the workflow job:

- strategy
  - matrix
    - os: 
      - ubuntu-latest
      - macOS-latest
    - arch:
      - x86
      - arm64
- runs-on
  - ${{ matrix.os }}
  - ${{ matrix.arch }}

However, once queued, the jobs simply wait indefinitely, insisting that no suitable runners are available.

Based on various sources, I have understood the following:

• For resolving the target image of each run, from the value of runs-on, operating system and architecture may be specified by labels that are evaluated independently. Each label may be given as an item in a list of all the labels, together restricting the attributes of the target runner image.
• ARM64 architecture is now supported within Github-hosted runners for Ubuntu.
• MacOS is available for both ARM and x86 machines.

As the documentation lacks clarity and consistency, I have experimented with several further variations of the architecture labels, including x86_64, x64 and arm, but in all case the results are equally disappointing.

I hope someone could identify my errors, and could help me understand the optimal usage of the workflow feature.

I wonder whether any document in particular accurately represents the complete and currently set of labels, and combinations of labels, within the pool of site hosted runners. It appears that much of the documentation, unfortunately, is dated or simply erroneous.

[doncjohn]

The part that is tripping you up is that the architecture labels are not generic for GitHub-hosted runners.

For self-hosted runners, a label set such as [self-hosted, linux, ARM64] can work because those labels are assigned to your own runner. For GitHub-hosted runners, you generally need to use one of GitHub's published runner image labels rather than composing ubuntu-latest plus arm64 yourself.

For example, I would model the matrix with explicit runner labels:

strategy:
  matrix:
    include:
      - name: ubuntu-x64
        runner: ubuntu-latest
      - name: ubuntu-arm64
        runner: ubuntu-24.04-arm
      - name: macos-arm64
        runner: macos-latest
      - name: macos-intel
        runner: macos-15-intel

runs-on: ${{ matrix.runner }}

A few notes:

• ubuntu-latest is currently the standard x64 Ubuntu hosted runner, not a base label you can combine with arm64.
• For Ubuntu ARM hosted runners, use labels such as ubuntu-24.04-arm or ubuntu-22.04-arm.
• For macOS, the current labels are split differently: macos-latest, macos-14, macos-15, etc. are Apple Silicon/arm64, while Intel uses labels such as macos-15-intel.
• x86, x86_64, x64, arm, and arm64 by themselves are not enough to select a standard GitHub-hosted runner.

So the indefinite queue is expected: GitHub is looking for a hosted runner that has every label in your list, and no standard hosted runner matches ubuntu-latest + arm64.

The most reliable source is GitHub's hosted runner reference table. I would build your matrix from the exact labels in that table rather than trying to derive them from OS and architecture separately.

[github-actions[bot]]

🕒 Discussion Activity Reminder 🕒

This Discussion has been labeled as dormant by an automated system for having no activity in the last 60 days. Please consider one the following actions:

1️⃣ Close as Out of Date: If the topic is no longer relevant, close the Discussion as out of date at the bottom of the page.

2️⃣ Provide More Information: Share additional details or context — or let the community know if you've found a solution on your own.

3️⃣ Mark a Reply as Answer: If your question has been answered by a reply, mark the most helpful reply as the solution.

Note: This dormant notification will only apply to Discussions with the Question label. To learn more, see our recent announcement.

Thank you for helping bring this Discussion to a resolution! 💬