Restarting Unhealthy Containers

[lig]

There was this one about 2 years ago on that topic #4826

It looks like Docker Compose is pointing to Docker and Docker is pointing to orchestration tools.

Is it possible to have restart policy to depend on health status? It looks like the job for Docker Compose for me.

Also, it was there for version 2 compose files in depends_on section.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically closed because it had not recent activity during the stale period.

[InsOpDe]

Please open again as it is still relevant

[pbelskiy]

When it will be able to?

[PhilippHomann]

Imho there is no way to do this in compose, because compose is not running during runtime of a container or services (except you omit -d on up).
To monitor the health state of a container and do actions on change you need to have a daemon.
Implementing those actions into docker daemon looks much smarter to me.
moby/moby#28400

[t3chn0m4g3]

Dirty, but works fine if docker-compose uses restart=always:

STOPSIGNAL SIGINT
HEALTHCHECK CMD if [ $(ps -C mpv -p 1 -o %cpu | tail -n 1 | cut -f 1 -d ".") -gt 99 ]; then kill -2 1; else exit 0; fi

• -p 1 Monitors PID 1 only and kills process if process CPU usage is > 99%.