Multiple Threshold values for a single parameter/metric using isPI Metric for Performance



I have gone through multiple blogs and seems I can't set multiple thresholds for a single metric. For example, my customer wants to have a warning alert at 60 and critical at 75% CPU utilisation.


There must be a workaround for this though. NNMi, being a standard monitoring tool, is expected to fulfill such a requirement.


Can you tell me what best thing can I do for my client for this requirement?




  • Hello Aditi,



    You can configure two Custom Pollers – one for Warning with threshold 60% and one for Critical with threshold of 75% CPU Utilization. Related to the two Custom Pollers you should configure two enrichments of the Management event if you want to change the node status, the message format and etc. Also, you could correlated the Warning incident as child of the



    Best Regards

  • Verified Answer

    • Hello Aditi,


    This can be also done without Custom Poller just by creating a few enrichments for the incident CpuOutOfRangeOrMalfunctioning. Lets say a user would like to have a Warning severity if CPU is in the range 60-69,

    Minor if in the range 70-79, and Critical if CPU >=80.

    • Set the threshold to 60 and create two enrichments with the following payload filters

    (ciaName like cia.thresholdMeasuredValue.* AND ciaValue between 60 AND 69)

    (ciaName like cia.thresholdMeasuredValue.* AND ciaValue between 70 AND 79)


    Change the severity in each enrichment to Warning and Minor.




    Thanks Sergey.


    Excellent!!! Seems to be working.


    However, I did a setting at payload giving ciavalue>=75 and the enrichment gets applied to all except the value 100. Strange it is. As a workaround, I have set another enrichment for value 100 and now want to test it.


    Will keep you updated.




  • The reson it does not work for 100 is because the custom attributes are strings (not numbers), so it does lexicographical comparison instead of numerical.

  • Hi Sergey,

           Assume utilization reached to 61% and alert got triggered, in next polling interval if it crosses 70% OR similarly 80%

    - Does it changes the SEVERITY in the same alert? 

    NO, bcoz MINOR alert raised when value between 70 - 79 will neither get closed nor changes when SEVERITY crosses >80

    This solution works only in the first iteration...