DX NetOps

 View Only
  • 1.  Physical Memory threshold

    Posted Jan 27, 2017 05:16 AM

    Hi 

     

    As of now we are monitoring Physical Memory on all over Windows servers at threshold of 85% utilisation and it is working perfectly fine by alarming when ever the physical memory utilisation goes above 85%. But now our management want to modify the threshold in such a way that they need the alarm only when the physical memory utilisation is above 85% continuously for 15 minutes in all the servers. Is there any way we can do it (we don’t want to change the server polling intervals as it will affect other monitors like up/down, CPU, other drive alerts). Can some one guide us on this?



  • 2.  Re: Physical Memory threshold

    Posted Jan 27, 2017 06:15 AM

    Hello, How are you monitoring you servers? By UIM integration, SNMP, or by Spectrum tresholds?



  • 3.  Re: Physical Memory threshold

    Broadcom Employee
    Posted Jan 27, 2017 07:27 AM
      |   view attached

    If you’re using Spectrum, highlight all of the models you want to change the threshold on (using List or Locater), right click -> Utilities -> Attribute Editor.  Open the Thresholds -> Device Thresholds and move the % Memory Utilization Threshold Duration (sec) over to the right hand side.

     

    Change the value to 900 and click Apply

     

    Cheers

    Jay



  • 4.  Re: Physical Memory threshold

    Posted Jan 27, 2017 07:29 AM

    Hi Renato,

     

    servers are discovered with SNMP and for configuring physical memory we go to server Information - system reasources - File Systems - File Systems (RFC 2790).

     

     



  • 5.  Re: Physical Memory threshold

    Posted Jan 27, 2017 10:07 AM

    I take a look on the docs and didn't find a way to do that for File Systems monitoring, and I think Jason's recommendation may solve your problem, only take a look at source on thresholds ( as image bellow ) to see if the monitoring data is the same, but if you have other products like UIM and eHealth I recommend you to migrate perfomance measures to then because you will gain a lot of resources of monitoring like baseline analysis and time over threshold is very simpler to configure