From time to time, the backend needs a restart because it shows a permanent 502 error: ctrl+C and ./server does the job.

It can be annoying since it human operation has to be done even if the server does not seem to be overcrowded.

Here are the log of one instance that needed to be restarted, for inspection.

Nothing anormal has been detected in this log (to be confirmed with another eye). 202312211010.log