I’m the sysadmin for beehaw.org and our user-base has almost tripled in 24 hours. Our site has been crashing so much over this time period.
We do have some volunteers that are trying to figure out a reasonable solution for when this happens again four weeks from now.
Do you have any recommendations?
Its a kludge, but in the time being you can also add a
docker-compose restartin your server cron every day or few hours. Lemmy restarts are pretty seemless, as the websockets automatically reconnect.The next release will have websockets removed which should fix some of the stability and memory issues.
Thanks!
No probs.
Are they being replaced, or is it just going to be all REST?
HTTP / rest client only.
/me goes and disable WSS support from nginx… :P
Some of the discussion of it is here. I was staunchly pro websocket, but it’s become too much work to maintain, and doesn’t scale well.