DX Unified Infrastructure Management

  • 1.  Monitoing the Nimbus monitoring

    Posted Oct 02, 2008 10:44 PM
    Hi,

    Maybe a strange title but currently I'm experiencing problems with my Nimbus dashboard. Sometimes it doesn't open the dashboard and in other cases all alarms are green while I know some of them should be red.

     

    So I was wondering if there are best practices to monitor if the dashboard is working and working normal. I can put a url_response in place to see if the link to the dashboard is fine but not if the dashboard opens and works fine. The same goes for emailalerts. If alerts are not send it doesn't mean everything is working fine.

     

    Many thanks in advance for your ideas.

     

    Wouter


  • 2.  Monitoing the Nimbus monitoring
    Best Answer

    Posted Oct 03, 2008 02:16 AM
    Wouter,

    I do not think we have encountered that kind of problem with our dashboards before.  We used to have an issue where it would be common to load a Web dashboard and find that it contained no QoS data (although I think alarms were still working).  Restarting the variable_server would fix the issue, and I think I figured out the problem would only happen if the hub had restarted or something like that.  I created a cron job to stop and start the variable server every morning just to make the problem less common.  That actually worked pretty well.

    Even though this is not the issue you are currently facing, it might be worth considering if you see a problem with QoS data.  Sorry that I cannot shed any light on your alarm problem.

    Regards,
    Keith


  • 3.  Monitoing the Nimbus monitoring

    Posted Oct 03, 2008 02:20 AM
    You could try turning up the log level on the dashboard server probe, and then using the logmon probe to check that log file for a specific log entry that might be indicative of the problem.

    You could also maybe edit the probe so it restarts itself at a quiet time. While it won't sort the root cause it may reduce the frequency of the problem.

    As an aside, Nimsoft are currently investigate an issue with our dashboard system were the dashboard does not update the correct status colour even though there are alarms existing. e.g at the top level an linked dashboard may show green but when you drill down something on there is showing a major alarm. This is despite enabling the propagation of alarms. Usually a dashboard probe restart sorts this.



  • 4.  Monitoing the Nimbus monitoring

    Posted Oct 14, 2008 10:30 PM
    I ran into the problem of alarm status not propagating when I built up a NimBUS test environment, but the problem somehow cured itself when I did a clean install on production hardware. So glad this wasn't just a problem with my setup and something that Nimsoft are looking into.

    It only seemed to show up with Linked Dashboards - if you had one dashboard with many panels underneath it, alerts would propagate up correctly.