I guess this could be done as an action on the charm to perform the restart and clearup log files; the replication failure should be detected in the update-status calls, (Unit is out of sync message), which would be a potential call to action to investigate and run the action.
I guess this could be done as an action on the charm to perform the restart and clearup log files; the replication failure should be detected in the update-status calls, (Unit is out of sync message), which would be a potential call to action to investigate and run the action.