Killing the cluster from the AWS ECS Web console, may not actually kill all the cluster’s resources. If you do it this way assuming it will, you may have a nasty surprise! To remedy the situation you need to:
The better way to kill the cluster is something like:
ecs-cli down --force,
but that might well fail if you have use a load balancer.
Docker in production has a history of "gotchas". The AWS interface to containers has a few too. Lets hope we can document them so things can improve.
Devops at Spuul. Any tips or suggestions? Reach out!