I want to know what happens in fargate if a container exceeds its soft / hard memory limits. As part of this I want to know would the container be killed? Would processes inside the container be killed?
I have a container service which has a very spiky memory profile, most of the time will be dormant however when it is used the memory spikes are of unknown size. These spikes are caused by large numbers of python processes being created (100+). These processes are making API calls followed by some computation once calls are complete. I want to know if fargate is a good use case for this.
Ideally I am hoping fargate can handle a job like this with no issues (i.e. not killing the container or processes inside the container).
Answer to the question above in writing plus any follow on questions that arise out of the original answer.