[ECS] [request]: Extend max container definition stopTimeout

[bennettgoble]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Tell us about your request
ECS container definitions allow the time to wait before a container is forcefully killed to be customized using the stopTimeout container definition property. This property has a default value of 30 seconds, and a maximum documented value of 120. AWS API Reference

It would be useful if the maximum value could be extended, so that services that are slow to shutdown can use stopTimeout instead of customizing the ecs-agent configuration option ECS_CONTAINER_STOP_TIMEOUT.

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?
Some containers may take longer than two minutes to shut down as part of expected behavior. As an example, the virtual world Second Life hosts many of its services using ECS, some of which require up to 30 minutes to gracefully shut down.

Are you currently working around this issue?
This problem can be worked around by providing ECS_CONTAINER_STOP_TIMEOUT to the ecs agent through /etc/ecs/ecs.config (linux.) However, this means that updating the value requires an instance refresh or similar operational chore to restart ecs-agent on impacted hosts. It would be easier if this value could be updated through task definitions without the current 120 second limit.

[rgoltz]

adding reference to issue #1020

[matt-domsch-sp]

See also #1020 (comment)

[joshuabalduff]

I noticed that this issue is over a year old and was wondering if there's an estimate of when this might get released if it is being worked on.