HA has become an integral part of the health and security of my home. On infrequent occasions, whether due to hardware or software glitches, HA will become unresponsive. In my case, this affects critical systems such as security monitoring, cameras, HVAC, and sump water level monitoring.
Considering the aforementioned, it would be great to run multiple instances of HA on separate hardware so that if one of the instances fails, it automatically flips to the available node in the cluster.
Ultimately, all instances would register with Nabu Casa, report on availability, and automatically proxy to the available node. As a bonus, upgrades could be done in a rolling fashion to avoid downtime during reboots and allow rollbacks when updates fail.