3 Replies Latest reply on Jun 23, 2008 10:41 AM by wildblue

    Ping latency

      ipMonitor 8.5


      I want to trigger an alert when the latency exceeds a certain threshold.  Am I looking right over it?  The ping monitor only fails when it gets no response.

        • Re: Ping latency
          mark wiggans

          Yes. To some degree this can be
          accomplished.   A “ping” monitor can be adjusted to fail at any time increment
          you choose.   This is done by making a combination of adjustments in the
          properties of the monitor itself.   In the “Timing” parameters, you will want to
          adjust the “Maximum Test Duration” from its default value of 300 seconds to
          something more aggressive say in the 10 - 30 seconds range.  Next you should
          look at the Notification control section and specifically at the value set for
          “Accumulated Failures per Alert”.   This value is set to “3” by default,
          however, if you wish hasten the alert notification process you will want to set
          this to a value of “1”.  This will cause the test to occur just once in the
          “Warn” state and then drop to a fail state.   Please note that ipMonitor is only
          able to trigger alerts from a “failed” state and can never do so from a “Warn”
          state.


           


          I hope this is
          helpful.

          • Re: Ping latency
            simonpt

            Good point, wildblue.  I'm seeing the same issue in 9.0.1.  The dashboard's 'Top 10 Devices by Ping' shows each device's response time but you can't seem to set any thresholding or alarms on response time (or latency).  You could have a device that is returning 800ms ping responses but you wouldn't know it unless you were glancing at the dashboard.  It'll always be green so you have to be vigilant to spot the excessive times.


             Rgds, Simon

              • Re: Ping latency

                I think I'm just going to try to create a little script that returns an error code of the average is > 100ms or so and run it as external process


                Unless someone already has something like that handy...