Given an ecs agent, when requested the agent is provisioned within the ECS cluster and seems to connect but does not get removed.

From what I can tell these agents don't actually get to the point they can be used within the pipeline, the jobs will keep requesting an agent and eventually will succeed leaving several dead agents around - I've attached the FINEST logs from the agent side 8dff8e99-fdda-4147-9879-fe71e026e5cd.log