The docs suggest to backup the `/var/lib/incus` directory (I am only using dir storage driver). Is it fine to run the backup while instances are up and in use? Is it better to use something like restic to backup the directory? Is it 100% reliable that I will be able to restore all profiles, network bridges, etc?
Lastly, what do you guys use for full backup of multiple incus servers in non-cluster?
I do a full backup of the host machines and export the instances as tarballs and the configuration as YAML.
If there is an issue with a host, I usually just redeploy it instead of restoring from a full backup.
If you have several non-clustered hosts then you can still copy instances and volumes between them. This is a simple way of doing the backups but you will need at least one of the hosts working at all times so that you can use it to restore the others.