DX Application Performance Management

  • Private Community
 View Only

  • 1.  Harvest Duration spikes coincide with Calculator Harvest Time spikes, these spikes occur every 10 minutes

    Posted Jul 14, 2014 05:03 PM

    I am running version 9.5 with 1 MoM and 7 Collectors. I am having high Harvest Duration, high Calculator Harvest Time, and high EM CPU % Used spikes that appear every 10 minutes. I'm concerned because these spikes can be anywhere between 8500 ms to over 20000 ms. The condition of having high Harvest Duration and high Calculator Harvest Time spikes typically only occurs on 3 or 4 of the 7 Collectors I have in my system and it's not always the same Collectors either. Has anyone else experienced this problem before?



  • 2.  Re: Harvest Duration spikes coincide with Calculator Harvest Time spikes, these spikes occur every 10 minutes

    Broadcom Employee
    Posted Jul 14, 2014 05:14 PM

    Are your EMs on physical or virtual hardware?

    What is the exact version of 9.5.x you're running?



  • 3.  Re: Harvest Duration spikes coincide with Calculator Harvest Time spikes, these spikes occur every 10 minutes

    Posted Jul 14, 2014 05:18 PM

    The MoM is on a virtual machine. Two of the Collectors are on physical machines but the rest of them are on virtual machines. The Harvest Duration + Calculator Harvest Time + EM CPU % Used spikes have occurred on physical and virtual machines. The version we are running is 9.5.0.



  • 4.  Re: Harvest Duration spikes coincide with Calculator Harvest Time spikes, these spikes occur every 10 minutes

    Broadcom Employee
    Posted Jul 15, 2014 12:56 PM

    On the virtual hosts, how many vCPUs are allocated and what are the CPU reservations set at?

     

    For all machines, how much physical memory does each host have?



  • 5.  Re: Harvest Duration spikes coincide with Calculator Harvest Time spikes, these spikes occur every 10 minutes
    Best Answer

    Posted Jul 15, 2014 10:24 PM

    Hi Tom,

    Besides looking at the VMWare environment configuration and making sure it's setup per our requirements(See APM Sizing and Performance Guide for details) key clue here is 10 minute interval between spikes,what you need to find out first is what's triggering that, are you running any Java Script Calculators\script on MOM or some sort of CLW query at same interval ?  There are some known high Harvest Duration spike issues (usually happen every hour out of the box) that can occur when SOA Performance Manager extension is deployed or when SAN is used (every 5 minutes) but nothing similar. You might want to review below threads for more troubleshooting tips

    -Enterprise Manager Performance Tips Interactive UI (https://content.plateau.com/nocache-content/ca/extended/product/apm/index.html)

    -Top 10 Performance issues on EM (CA Tuesday Tip: Top 10 common causes for EM performance issues - Checklist)

     

    If nothing helps then I would recommend opening a support issue with logs from all EM's and screenshots of relevant metrics so we can help you further.

    Regards,

    Kulbir.