I have been trying to get a few PRs merged since yesterday with no luck. My sense is this is not because of any specific problem, either with CI or with the code, but more a result of the general instability of CI for the k8s collection.

In this case, the runs are frequently failing at the same task: https://github.com/ansible-collections/kubernetes.core/blob/73499d9a0960835a3ed340286e7c7039d99fdc21/molecule/default/tasks/apply.yml#L295. This task will sometimes pass on some jobs, but there's always at least one job that fails. I don't know why this task has decided to start failing now. As far as I can tell, there's nothing that has changed that would cause it.

I'm not sure when this issue can be called done. I assume eventually the PRs will start to pass and we can merge them. This is exactly the kind of problem we have been dealing with almost every week for months. Just trying to document it here.