I have a relatively complex workflow, but random jobs within it take 2-9 minutes to spin up with no visibilty into what’s taking so long. These were from the same commit, taking a workflow that should have taken 2 minutes, to one that is going to take at least 14 when it finishes.
(the infuriating thing is that this job with 5x parallelism is done, in the time this fifth one spun up the other 4 did all of the work and finished in under 20 seconds, leaving this vestigial container to sit there, “spinning”, blocking the rest of the workflow)
Is there anything I can do to fix this?