Oof, what a ride!
I finally forced myself to do some upgrade and maintenance tasks on the server (the actual system, not just individual services when inevitable).

Gallery software #Piwigo => plain sailing, even less issues than anticipated to deal with
#NextCloud => pretty much the same, even though it needed several iterations due to me having been 2 (still supported) major versions behind
Wife's #Moodle => also major version upgrade, still no issues at all. A couple git commands, call the update script, all done and dusted
#Cubecoders' #AMP => yyeeeah... no, all good! One of the #Minecraft instances didn't completely upgrade automatically but giving it a slight manual poke also fixed that. Great!
#Akkoma => Oh right, that uses Erlang and since I installed via git instead of a binary distribution I need to run the build scripts again. Shouldn't be too bad, the upgrade instructions are fairly straightforward. AS LONG AS YOU DON'T RUN INTO PROBLEMS! 😨

Idiot me - against knowing better - was too lazy to make a backup before attempting the upgrade. Result was crashes right on service startup. Something with mismatching keys and stuff. Did I already mention I'm an idiot? Well in case I didn't: I'M AN IDIOT!
Okay, time to straighten up, no use in moaning over spilled milk. Best way to approach this problem? First try to get #Ubuntu up to date so I won't run into more issues than necessary should I need to install stuff for rectifying the Akkoma issue. On login the MOTD warned about having reached EOL for this release... Oh. Yeah right, I upgraded to a non-LTS release at one point to satisfy Akkoma's dependencies. Time to learn the lesson that those releases only get 9 months of support and after that any onboard means for dist-upgrades simply fail unless you let #Shuttleworth into your wallet for some superduper subscription plan for stuff that's freely available on the internet as #opensource. How convenient!

[== Fastforward through roughly 3h of teaching new indecent vocabulary to neighbours but eventually succeeding in getting the system to update without sending money to a South African millionaire while thinking about my options should the instance turn out to not be recoverable with my abilities ==]

Finally! Do a restart and then gather a status report.
[== Oh those long, long minutes till the SSHd becomes reachable again after the forced disconnect for reboot... Is everything going well? Should I preemptively pull up the management panel of the hosting provider? ==]
Yes, we're back! Pull up syslog to see what exactly Akkoma is complaining abo--
Huh? No errors, everything started flawlessly. Wait, even Akkoma?! *quickly checks browser* Yep, even Akkoma, as if nothing had ever happened 🀯

I think I need a therapeutic #whisky now.
Akkoma