I forgot to mention that the problem is happening only with my arm cpu node. The other node with an amd64 cpu is working as always.
It could be that it’s slow to respond on requests from the dashboard, but it is better to run the database check anyway, if you want to troubleshoot.
I have noticed this happening on mine in the past as well although I quickly identified MY issue was always high IO / disk latency although I would imagine this could also be caused by CPU utilization on lower power systems. I originally had nodes running in a raid pool (don’t do this) and when there’s a lot of disk IO (primarily due to database files) the dashboard lags. I’ve noticed it being anywhere from a few seconds to several minutes in my case but depending on your setup it could be much longer.
Garbage collection, filewalker and copy tasks if your moving nodes to a new location can all cause this. Additionally don’t use raid, make sure you don’t have SMR drives as this will be far worse with those, and make sure your nodes aren’t running on the same disk as the OS or other programs. If you still have issues you can try moving the database files to a good SSD.
EDIT:
Also, if your using custom dashboards that rely on exporters this creates additional IO especially if using log exporters. Export your logs to a different drive and / or reduce the frequency their scanned.
1 Like