Is it possible to ‘delay’ the dequeuing of a task that has failed previously? As a per job setting maybe?
Say, we have a job that connects to a server that gets intermittently busy - and thus times out sometimes. If we could delay the next dequeue of the same task by something like ~30 seconds, the task would go through just fine. But deadline being so fast, it usually accumulates a few failures within that time period already.