ci: support re-trying jobs in triggered cerbero sub-pipelines
When a job in a cerbero sub-pipeline fails, the cerbero trigger job fails.
It's then not possible to re-run that job in the sub-pipeline and have the parent pipeline succeed, one always has to spawn a new cerbero pipeline and rebuild everything from scratch (which takes about another hour give or take).
I wonder if it would be possible to make the trigger_cerbero_pipeline.py script ...
- first check if there are successful sub-pipelines already (e.g. from retrying a failed job on an existing sub-pipeline)
- next check if there are still-running sub-pipelines, and if so, wait on the last running one instead of triggering a new one
- only trigger a new sub-pipeline if there are no existing sub-pipelines or none that are successful or still in progress.