This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

MIB for CPU Utilization - Linux Server

Our server team has three Linux servers.   On one, NPM is showing high CPU utilization (100%).  The other two are showing more reasonable CPU utilizations (under 10%).

Our server team wants to know which MIB is NPM looking at to see these values.  Does anyone know this?

Thank you.

Michael K

  • I believe theyare using:

    For servers (Windows, Unix, Linux, etc.), they use HOST-RESOURCES-MIB::hrProcessorLoad (1.3.6.1.2.1.25.3.3.1.2).

  • 2
    hrProcessorLoad
    INTEGER32
    Legal values: 0 .. 100
    ReadOnly
    The average, over the last minute, of the percentage
    of time that this processor was not idle.
    Implementations may approximate this one minute
    smoothing period if necessary.
    

    hrProcessorLoad  INTEGER32


    Legal values: 0 .. 100 

    ReadOnly The average, over the last minute, of the percentage
    of time that this processor was not idle.
    Implementations may approximate this one minute
    smoothing period if necessary.

    Do you have APM to where you could discover the processes on that system and see how the CPU load is being used?

  • How does this work for multi-processor systems? 

    I did a MibWalk on one of my systems and see an hrProcessorLoad for each CPU in the system and would like to know if and how this is taken into consideration?

  • This thread I found suggests that they are using something different...

    ?PageIndex=1

    Would be nice to get confirmation for sure on how this is being calculated because I am not sure if it's accurate or not but I am having doubts.

  • Yeah you're right byron... sean makes the unix server cpu calculation sound pretty complicated:

    Using the new OID ssCpuRawIdle:  1.3.6.1.4.1.2021.11.53.0

     Transform Formula

    Truncate({100-{ssCpuRawIdle}/({ssCpuRawKernel}+{ssCpuRawIdle}+{ssCpuRawSystem}+{ssCpuRawUser}+{ssCpuRawWait}+{ssCpuRawnice})*100},2)

     

    I too wonder what this means with multicore.... and hyperthreaded cpu's now???

  • If I am not mistaken NPM is doing pretty bad job on linux (net-snmp). NPM dosen't understand which unix/linux version? It doesn't understand it's CPU/DISK/Memory accurately. We need an urgent solution if possible?

  • If you use SAM you can get a lot more of this information.  Also with more recent versions NPM & SAM have become much better at handling Linux.  NPM still doesn't see the Linux version but that is less a failure of NPM and more a failure of the net-snmp agent which is terribly maintained.  Since NPM talks to the SNMP agent on a system it relies on the data that agent provides.  SAM gives you a lot more options when it comes to collecting data from a system so I would suggest you give that a try since it's more designed to do Server monitoring.

    I have a UnDP that will do Load Average for net-snmp systems such as Linux and Unix, you can find that HERE.

  • Hi Byron,

    If I am not mistaken you need an agent on the server side for SAM. And nobody wants to install a large sized agent on their servers. Also there are other NMS's (especially free ones) that will show the true information on linux boxes like OS information/CPU/Memory/Disk. So I think the real problem is Orion linux support. Is it too much to ask? Not all systems are windows these days. But thanks for your help anyway.

    I still need a workaround for all these information. emoticons_plain.png

  • SAM doesn't use agents.

    I realize not all systems are Windows.  We are a cloud hosting provider and have many Linux and even some Unix systems that we support and I have been able to get CPU, Disk, and Memory information for those systems just as easily as I can for our Windows systems.  In my experience Orion fully supports Linux, if you aren't receiving this information you may not have things configured correctly.