Open stephanie0x00 opened 10 months ago
Hm, This would mean some other service is using up too many postgres client slots or creating too many threads.
It looks like this is caused by rocky or at least partly caused by rocky. When you go the katalogus page, it will fetch all images and for each request we use the database to check authentication and authorisation.
When opening the 'Objects' page postgres temporarily times out while gathering all records for the page. This occurs when having a lot of objects (currently 27k objects). After waiting for some time it automatically resolves itself, however the user will think that one of the services might be down instead. This was found on the testserver.
Error logs are attached. postgres-crash-alexa1000.tar.gz
Screenshots
OpenKAT version commit 6d2f443f1eaedd1dfdc5735c2e44f34157b06adf (HEAD -> main, origin/main, origin/HEAD)