Run script on error

[ilovethensa]

It would allow comin to for example send a notification when the deployment fails

[nlewo]

There would also be two others ways to achieve this:

• comin exposes Prometheus metrics: an alertmanager could send an alert
• wrapping this switch-to-configuration script to detetch its failure and run an action on failure