sorry about that, have not noticed this in monitoring, nor in browser, otherwise i would have restarted lemmy-ui. i have done so now.

Sorry should be better now.

oh, sometimes lemmy-ui needs a restart, usually the first 504 from the monitoring would be the signal for me but i didn’t notice this in a while - i was actually just wondering about it. will check on it asap, matrix is unrelated

edit: restarted. and indeed here and there i am already deploying anubises - just not too happy to seemingly have to do one instance per domain - even if i can update them using watchtower and have them share a configuration file... even lost track a little about where i have put them already lol

sorry, i had noticed that the monitoring was triggered a little while ago which indicates that its time to restart the frontend to mitigate this issue, but i forgot

will look into an alternative frontend but not sure if that would see much use

i try to investigate but its a bit tricky - could you confirm that this does not happen in apps? from my tests it looks like its actually a problem with the lemmy-ui itself. not lemmy or another component, as when moments where i cannot curl the ui on server level, i still can curl the lemmy server for example

i sadly struggle gathering helpful logs but i am confident the cause will be found

interesting i primarely use an app and everything seems responsive. will try to investigate

first fork i noticed and following with careful amount of hope: https://github.com/we-promise/sure

i have added some headers (most interesting perhaps slice) - however arte currently times out when testing, aeharding does not seem to present opengraph stuff to test, peertube seemed to work normally before my adjustments

playing back an uploaded video didnt look like it was sliced however it played without showing any obvious errors anywhere

i have added proxy_hide_header Access-Control-Allow-Origin; which seems to help with that reflected origin