Ubuntu cron.hourly script to monitor and restart non-running Docker containers

docker-container-monitor-restart.sh

#!/bin/sh

#
# Check on "running" status of critical Docker containers
# and restart them if they are not running.
#
# This is necessary because after a hard host machine reboot, many Docker containers will
# will not start cleanly. Since Docker restart policies and healthchecks only take effect
# once a container has successully started up once, these failed containers will never
# start.
#

# install in /etc/cron.hourly/docker-container-monitor-restart 

criticalContainers="container1 container2 container3 container4"

for containerName in ${criticalContainers} ; do
  containerStatus="$(/usr/bin/docker inspect --format '{{.State.Status}}' "${containerName}")"
  if [ "${containerStatus}" != "running" ] ; then
    /usr/bin/logger -t cron.hourly \
      "Docker container ${containerName} is ${containerStatus}. Restarting ${containerName}..."
    /usr/bin/docker restart "${containerName}" | /usr/bin/logger -t cron.hourly
  fi
done