Curious hang after a long time.

Not sure how to reproduce this, but implement a regular broadcast by the daemon into MQTT that if not received for more than 2 minutes should reboot all ESP nodes as the MQTT link might have been dead.
Reconnecting/rebooting might fix it.