I was trying to be nice to my little NAS running this instance (and other instances and websites and email and *arr things and ...), by limiting the number of threads for inbox, outbox, backfill, etc to very low numbers.
Then I added every relay known to Cthulhu, causing the queue backlogs to get kinda stupid large. What the heck, I have NVME cache protecting the spinning rust in the NAS. Other than maybe running out of CPU due to i/o wait what could it hurt ... cranked the concurrency up quite a bit. Lo and behold, everything's just fine and processing is keeping up.
#sysadmin #selfhosting