which indicates the CPU percentage by CPU, but I could not find a reference where the percpu was set False to get the average CPU percentage use. So, it seems to calculate the average CPU utilization from per CPU usage, more in the windows way (all utilization rolls up to 100%) than, I guess, the Linux way (20 cores at 100% would result in an “average” of 2000%). I think, it does that around here …
and in that function executes a group by
I am wondering whether during that average system calc (upper graph) something goes wrong when the data from 20 cores are involved,
I would certainly hope so (the the average is wrong), because I’d be concerned if the CPU really were averaging 50% usage when the NAS is completely idle
nothing reading, nothing writing, no scrubs or balance, no services other than SMB enabled.
It’s been like that for a VERY LONG time… I rather ignore it… It’s just one of those low priority things in the GUI compared to all the huge work upgrading everything else I guess.
Oh, do you know if there is an open bug for it? I dont know ANYTHING about git or any of the bug tracking type stuff. but even if it is a low priority one, it would be good to have a record of it somewhere.
@McFaul I checked and could not find an open Issue on github for this, though I think I remember it has been brought up here on the forum a couple of times. In any case, I have opened a new issue on github for this:
In the meantime, you could use the Netdata rockon if you want to get a better view of CPU utilization (both system average as well as CPU level) as well as many other parameters of your box.