CI folks have come up with a convention for systems to indicate that a test is scheduled or running: we can publish a CI message and also send a 'result' to resultsdb, which is not really a result but an execution status. See https://resultsdb.fedoraproject.org/results/40676203 for an example - basically, you publish a result as normal but with the outcome as RUNNING or QUEUED.
RUNNING
QUEUED
We probably can and should do this, then we can also have Bodhi show when tests are queued/running which would be nice (see https://github.com/fedora-infra/bodhi/issues/4853 ).
Practical issues:
This is now partly implemented. We publish results and messages in cases where an openQA internal event happens, which is most simple cases.
However, there are definitely cases where there's no openQA internal event when a test gets started or restarted. https://progress.opensuse.org/issues/123625 is where I'm tracking this upstream; it's rather a difficult problem upstream, unfortunately. I am hoping to be able to patch the holes there as best I can when I get time to look at it again, I started on it a few weeks ago but have been nerdsniped by other stuff since.
I think I'll close this ticket as the basic work in fedora_openqa is done. As we patch holes upstream, that should automatically mean we publish messages/results in more situations downstream without needing to make any further specific changes to the fedora_openqa code.
Metadata Update from @adamwill: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
This issue has been migrated to Fedora Forge: https://forge.fedoraproject.org/quality/fedora_openqa/issues/101
Please continue any further discussion there.