A watchdog is introduced into the Core which should make sure that the system looks "sane", and if not try to take corrective action.
The first watchdog check introduced by this change is checking the number of managed nodes vs. the number of nodes in the configuration. If something looks off (they differ in the number), a reconfiguration is triggered. This is run every 5s.
|For Gerrit Dashboard: JVMCBC-1045
|JVMCBC-1045: Implement basic state Watchdog in Core