check_nt CPULOAD returns extremly high CPU Load above 100%

ferox147 commented 10 months ago

Issue and Steps to Reproduce

Running "check_nt" CPULOAD on a physically installed Windows Server 2022 Standard returns e.g.: CPU Load 18446744017345409802% (5 min average)

'./icinga2/addons/check_nt' '-H' '$hostadress$' '-l' '5,80,90' '-p' '1248' '-s' '$secret$' '-v' 'CPULOAD'

Expected Behavior

In fact, a correct value between 0 and 100 % should be returned.

Actual Behavior

Unfortunately, an extremely high, unrealistic value is returned, which does not reflect the behavior of the system. (18446744017345409802% | 5 min avg)

Details

NSClient++ version: NSClient++ 0.5.2.41 | 2018-04-26
OS and Version: Windows Server 2022 Standard
Checking from: Icinga
Checking with: check_nt

Additional Details

CPU: 2x 8-core / 16-threads: 32 logical processors

ferox147 commented 9 months ago

Also tested with check_nrpe command, I get the following negative percentage result:

OK: CPU load is ok.|'total 5m'=-3041937069%;80;90 'total 1m'=0%;80;90 'total 5s'=0%;80;90

Marco-Jacobs commented 4 months ago

It's the same here. We monitor around 100 hosts with NRPE. This behaviour only occurs sometimes and only on 5 hosts.

The NSC versions are different 0.5.x versions
The operating systems of the target computers are different Windows Server versions

dneuhaeuser commented 2 months ago

same here with one 2022 server, which has 56 cores.

Maybe related to core count? I don't see this problem on 2022 servers which have 16 or less cores.

v0.5.2.41 on all servers.

mickem / nscp