Configuration & Administration

Expand all | Collapse all

CPU REPORTING INCORRECT VALUE

  • 1.  CPU REPORTING INCORRECT VALUE

    Posted 07-31-2019 06:27 AM
    We are  getting a lot of (hundreds) notifications where the CPU usage is reporting incorrectly. 

    "Summary: threshold of cCPUUtilizationCrit exceeded: current value 5735.000000"

    These events are generated and flooded when we restart zencommand services.

    CC : 1.5.0
    RM:  
    v6.1.2

    Can someone help me identifying the root cause of this issue?


    ------------------------------
    Jagan

    ------------------------------


  • 2.  RE: CPU REPORTING INCORRECT VALUE

    Posted 08-01-2019 09:49 PM
    Hi Jagan,

    We have experienced the same behavior whereby after stopping the collectors or all of Zenoss for that matter, leaving it down for 15 minutes or longer, then bringing it up, we see false/impossible values for CPU and Interface utilization (greater than 100%). In all cases of this, the data points are defined as DERIVE. In working with Zenoss Support, we believe there is a COUNTER value stored before the collectors/Zenoss comes down. After the collectors/Zenoss comes up, Zenoss starts collecting the COUNTER values again. The difference is calculated between the two values but there could be several values missing between the two COUNTER values. The resulting derived value ends up being a higher and inaccurate value.

    Zenoss Support has opened a case to get this fixed. Until that happens, we disable triggers for High CPU and High Interface utilization after bringing Zenoss up and closely watch the Event console until there are no more being created. After that, we enable the triggers.

    HTH
    Dennis Stackley
    Systems Integration, Senior Advisor
    NTT DATA Services

    ------------------------------
    Dennis Stackley
    Systems Integration, Senior Advisor
    NTT DATA Services
    Divide CO
    7202551377
    ------------------------------