Is there any way for MaaS to have redundancy or high availability for the Region or Cluster controllers?
MAAS is designed with this in mind but is not quite there yet.
The region controller can be split up into several appservers and the postgres database can be similarly split. The only SPOF is the celery daemon that runs as part of the region, and we have plans to also split that and make it HA.
The clusters themselves are not designed to be individually HA because you just run multiple clusters and do it that way.