node_exporter cpu request

[rkojedzinszky]

What did you do?
Just observed, that node_exporter has 102m cpu request. What is the real reason behind 102m ?

Did you expect to see some different?
I expected to see a rounded value there, either 100m or 10m.

Environment
General.

[brancz]

No reason that I can remember. Happy to have it be an even number if you want to change it.

[omerlh]

I'm noticing that 102m is not enough for node_exporter, and it might need something around 800m or even more, am I the only one experiencing it?

[brancz]

wow, that is quite a lot, do you happen to have any special collectors running?

[omerlh]

Nope, using vanilla configuration as provided by kube-prometheus and latest version...

[rkojedzinszky]

Strange. On my pi boards node_exporters consume 0.02 cpu per node. Regards, 2020. február 3. 10:50:59 CET, Omer Levi Hevroni <notifications@github.com> írta:

…

Nope, using vanilla configuration as provided by kube-prometheus and latest version... -- You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: #214 (comment)

-- Ezt a levelet az Android készülékemről küldtem K-9 Mail-el. Elnézést a tömörségemért.

[omerlh]

I actually had to increase resources to solve throttling issue, so the discussion should move to kubernetes-monitoring/kubernetes-mixin#108

[paulfantom]

CPUThrottlingHigh is now only an info-level alert and should be inhibited if nothing else is firing. We have an issue in #861 tracking adding such inhibition.

I would advise increasing resource requirements only if you see that node_exporter scrapes are taking too long. In other cases, there is nothing to be worried about.