What is this?

This knowledgebase contains questions and answers about PRTG Network Monitor and network monitoring in general.

Learn more

PRTG Network Monitor

Intuitive to Use. Easy to manage.
More than 500,000 users rely on Paessler PRTG every day. Find out how you can reduce cost, increase QoS and ease planning, as well.

Free Download

Top Tags


View all Tags

My Probe goes high in CPU utlization, while the machine it runs on says 1%???

Votes:

0

I'm running the probes on Hyper-V machinery. My primary probe has begun alerting me of high CPU utilization (health sensor). As it runs as a Hyper-V I've gone into the Hyper-V host, and looked on the CPU load for the PRTG probe, and that reports 1-2% CPU load? We've changed a LOT of our WMI sensors to SNMP sensors to lessen the load, and make the checks more stable (we're not that thrilled with WMI), especially for our clustered environments we get better results with SNMP after a node switch.

cpu health prtg

Created on May 29, 2018 7:37:58 AM



7 Replies

Votes:

0

Hello fribse,

Sounds all very reasonable, you can't go wrong with ditching WMI sensors and replacing them with SNMP sensors where possible.

Most likely the Hyper-V host does have its own way of determining the cpu load of a vm and it's different from what the health sensor of the probe does, which does some lokal WMI query towards rge VM's operating to retrieve cpu load, which I don't think is what the hyper-v host does as it is able to show cpu load of the machine no matter what os it runs, without the need of having Windows credentials, SNMP or SSH credentials as required in PRTG where you always run a query towards an os. See also this article I'Ve stumbled upon.

To get furthermore another "opinion", you can add additionally a Windows CPU Load and an SNMP CPU Load sensor and see what they report.

Also quite of interest would be to double-check with the local task manager on the vm to see what this one has to say.

Kind regards,

Erhard

Created on May 29, 2018 1:29:29 PM by  Erhard Mikulik [Paessler Support]



Votes:

0

I can see that the local taskmanager for the probe goes to 100%, and is generally very high. It's setup as a 3 core machine, with 8 GB rams. We currently have about 3000 sensors.

Created on May 29, 2018 1:48:45 PM



Votes:

0

So.....whom do we believe now? The VMs operating system or what Hyper-V reports for it? :) Keep in mind, Hyper-V does determine the load for the VM differently. Or at the end of the day it's some Hyper-V bug :)

Created on May 29, 2018 2:10:44 PM by  Erhard Mikulik [Paessler Support]



Votes:

0

So, how to solve it, more ram or more cpu?

Created on May 29, 2018 2:14:14 PM



Votes:

0

What processes do you see in Task Manager that cause the high load?

Created on May 29, 2018 2:41:42 PM by  Erhard Mikulik [Paessler Support]



Votes:

0

I've gone with a 6-CPU setup now, and I've adjusted the interval on some of the sensors, it seems to settle to a more reasonable level, and the RAM does not max out, it uses around 5 GB of the 8

Created on May 29, 2018 2:42:22 PM



Votes:

0

Ok, if you want me to have another look at logs etc., please send me an email ([email protected]) or open a ticket through "Setup | Contact Support" referencing this thread.

Created on May 29, 2018 2:51:41 PM by  Erhard Mikulik [Paessler Support]




Disclaimer: The information in the Paessler Knowledge Base comes without warranty of any kind. Use at your own risk. Before applying any instructions please exercise proper system administrator housekeeping. You must make sure that a proper backup of all your data is available.