Before this the last change I made was adding a POP account. The server didn't have enabled POP access so Indigo reported it in the log every minute. But it was when the message ReceivedConnRequestTCP_ts() caught exception: SocketError -- socket accept error appeared that it would need to be restarted.
I found this other thread with a similar error but it wasn't solved. It did mention something about email briefly near the end reenforcing my suspicion it was related to that so I enabled POP in the email server and verified it that it worked. Restarted Indigo Server and confirmed that the errors cleared. It didn't become unresponsive anymore after that either.
After a few hours I checked the server and noticed it was rather clumsy--slow. There were no errors in Indigo's log. Then I noticed Indigo was using 200% of CPU. It would spike to that usage about a minute or less after starting. This is the server process, not the plugin host processes nor the client process. Regardless, I started disabling plugins to rule them out until I got inpatient and just stopped the server to disable them all at once but it made no difference when I started back the server. About a minute after, it would spike the CPU to 200% again and hold it there.
I switched focus and I upgraded the OS from High Sierra to Mojave — if you can that an upgrade — but it wasn't fixed. I tried plugging in the PowerLinc modem in another USB port of the hub-workaround-thing it's on and changed the hub itself to another port on the Mac as well but nothing improved the situation.
What causes this? Is it well-known? Could something like a corrupt database be the culprit--I ask because I'm OK wiping the computer or moving it to another computer and wiping that too, but I'm not looking forward to manually enter my whole database manually if it has become damaged and I need to start from scratch. :/
Hopefully it's a known thing.
Indigo Pro 7.0.4 on macOS 10.13.6->10.14.6.