@stgraber I have experienced an issue with Incus v6.18. After the first three Incus v6.18 updates my NPM container started going offline despite no errors in the NPM docker container which is nested in and incus container and the application being online from its perspective. The GUI was no longer accessible and reverse proxies were going up and down every few seconds.
A known working backup of the Incus container had similar problems and I tried three different older backups with the same result. A new installation of NPM also had the same issue. The server in question hosts 33 other containers and they all appeared healthy. For that matter, NPM appeared healthy, but was clearly misbehaving.
The incus server also had no remarkable log events. Finally, I rebooted the incus server and the problem was gone. This also happened with a previous version of incus where there were multiple updates to the same point version. I didn’t report it at the time because I felt it might be a one-off.
This leads me to believe that a best practice might be to include a reboot after repeated patch versions of the incus server although not indicated. Can you comment and have you heard of anything like this occurring?