implement auto-healing function as part of zebrunner.sh

[vdelendik]

aswe gonna to limit number of automatic retries: #95 we need smarter method for reanimating devices including full container re-creation: think about naming ./zebrunner.sh doctor.
This option should be added as part of hourly cron during setup
It should

1. scan every registered device
2. detect containers state
3. skip all disconnected and healthy
4. recreate container(s) for unhealthy
5. analyze logs and state so devices are fine
6. if something wrong and auto-healing doesn't work - down device containers

[vdelendik]

removed milestone as previous 2.2.x releases are pretty stable and reconnect devices individually. let's analyze upcoming 2.3 and close as won't fix if no complains.