I'm not sure what to do about this, but the current behavior of our GHA scripts, to abort any already running tests on the same branch when a new update is pushed, is not helpful.

Often, the end result is that you get no results at all, since the previous test run is partially through, but is then just aborted, and the same will happen with the new one that replaces it, and so on.

I think a better behavior is to let running tasks finish, but don't start any new. If there are multiple runs queued up, everything in the queue prior to the latest version, should be removed.

That way we will not end up with a long stack of queued runs, but we will not either abort an ongoing run, and wasting all the time that has already been spent on that run.

A perhaps even better solutions would be to cancel a task if it has spent less than X minutes running, but keep it and let it finish if it has spent more than X minutes.

I have no clue if any of the above ideas is at all possible to implement in GHA. I guess not. But we can likely make this better than it currently is, though.