It’s always good to red the docs, but I often skip myself :)

They have this nifty tool called pve8to9 that you could run before upgrading, to check if everything is healthy.

I have 3 node cluster, so I usually migrate my VMs to a different node and do my maintenance then, with minimal risks.