[ECS]:backoff on container restarts

**Tell us about your request**
I want the restart behaviour of the ECS-agent to be configurable so that it's possible to exponentially back-off and increase the time between restart requests.

**Which service(s) is this request for?**
ECS

**Tell us about the problem you're trying to solve. What are you trying to do, and why is it hard?**
If a container crashes, ECS will immediately turn-around and launch a replacement.
This will lead to massive calls to some AWS services from the ECS service or agent, this is generating massive amount of CloudTrail logs and associated costs.

There is currently no control over this behaviour in ECS

**Are you currently working around this issue?**
Not possible currently, other than to make sure that containers are never crashing

The issue concerns ECS-agent container restart behavior but names no repository files, tests, or implementation entry points. Start by identifying where ECS or the agent currently requests replacement containers, then check whether an existing restart-policy configuration can support exponential backoff; done means a documented, configurable behavior with defined timing and limits.

Written by the indexing model from the issue text.