After setting up a job to use fargate, it initially worked well.   I came back the next day to find 100+ jobs queued and the console spewing messages about ecs rate limit.

This likely resulted in huge log files... I would have expected the jobs to fail, but they kept trying to reschedule.   Not cool.