https://www.youtube.com/watch?v=R_NAoNd4YyY
because ECS and Batch are not so great
simple tool to kill ECS instances that are misbehaving, purge a batch job queue, and replay failed jobs.
Usage: sovereign [-k] [-p reason] [-s status] queueName
-k Kill the instance with failures. Requires appropriate aws keys and permissions
-a ask to retry all failed jobs
-p Purge the queue with the specified message. Conflicts with -k or -s
-s Which status to look at. Must be one of ${JobStatus.values().mkString(",")}. Defaults to ${JobStatus.FAILED}
-h this message
-r replay failed jobs that failed due to container error