Which Logs Relate to the GUI?

[jwsl224]

i have this issue with a server loosing connection to the GUI under certain workloads. or, the connection getting unstable. the VM's will for the most part keep running, though some might hang after a long time. i'm guessing it could be some kind of driver issue. i tried

journalctl | grep timeout

but didn't get much interesting info. what search term could get me info on the GUI? or, maybe whatever maintains the connection to the cluster for that node?

 

[fba]

The service for exposing API and providing GUI-Access is pveproxy. journalctl -u pveproxy.service will show you all log entries of this unit.
You mentioned certain workloads, could it be that the network connection itself is saturated?
From the physical cores of your server, how many are not used by VMs? In other words: is one always free and available for PVE?