====== Overview ======

It is possible that due to unattended updates, a docker container is shutdown because dockerd was restarted. Without a restart policy, it remains “dead” and it is not started again. We encountered this problem at dev02. Thus, the policy should be provided in the run part:

<code>
docker run --restart always [..]
</code>
Documentation: “**always** Always restart the container if it stops. If it is manually stopped, it is restarted only when Docker daemon restarts or the container itself is manually restarted.”


===== Logging =====

To limit the logging, use the following parameters:

<code>
docker run --log-opt max-size=50m --log-opt max-file=3 [..]
</code>
To see if a running container has this log options, use

<code>
docker inspect name_of_container
</code>
and search for LogConfig


===== Health Check / Ping =====

In the Dockerfile, add something like

<code>
HEALTHCHECK --interval=5m CMD curl http://localhost:8000/health || exit 1''
</code>
The CMD has to return exit code 0 for OK and exit code 1 for UNHEALTHY

Then ''%%docker ps%%'' will show something like this:\\
Up 15 minutes (healthy)\\


Keywords: automatic restart dockerd container ping monitoring health check


===== Optimization =====


  * Consider ''%%–compress%%'' for docker build
  * Each RUN creates a new layer → stack your RUN commands