The process went through smoothly. I have also used the opportunity to split up a singular lemmy container into individual tasks - this has enabled a seemless upgrade process with no downtime, bar a few process quirks I need to work out.
There have been some federation fixes merged into this release, so the situation should definitely be improving overall!
I will make a more detailed write up of the whole setup later on, other admins might find it useful. Or not.
Try enabling debug log for federation - helped for me. Phiresky guessed it's due to federation queries slowing down enough to not get stuck.