[srd]

That depends on the implementation of the queue. Most queues have a "failed/retry" concept, where they mark failed jobs to be retried in some set future point in time. So one failed job does not hold up your entire queue processing.