Folks;
We're using AWS Glue to drive a number of ETL type tasks ... one of which needs to be run in a Container because of some legacy POSIX dependencies. For this, I have a short Python / boto3 script to run_task() from Glue just so that we can centrally manage all workflows.
I can run my ECS task (has to be EC2, not Fargate), no problem. And FYI, I am using waiters to watch for provisioning then running. My Glue hosted stub exits only after the ECS task completes.
But I cannot figure-out how to get the exit status of the run within the Container. A bonus would be to also to have the ability to capture stdour/stderr from ECS via this Python stub.
Looks like I / we could query CloudWatch via boto3 after the ECS task completes, but this seems like a cumbersome approach (and I don't want to do this because the API rate limits for 'describe log') ...
I am hoping for pointers to where I may have a gap in understanding, or confirmation that this cannot be done today.