2 Replies Latest reply on Oct 18, 2012 1:54 AM by matthi77

    Hardware Monitoring Issues with Juniper Devices




      Thank you very much for this great new hardware monitoring feature!! It´s very useful for us!!


      However I have noticed some problems in combination with Juniper Devices:


      1) Components in "Standby" status (for example failover routing engines) remain in a permanent warning state.Shoudn´t they be marked as up, since they do not have any problems...

      2) EX4200 Switches show up with 3 fans and every fan is titled as "Fan 1". Would be simpler to distinguish with a clear numeration of the fans...


      One other question arised: Is there a possibility to adapt hardware monitoring thresholds  (for example fan speed warning and error thresholds; temperature thresholds?). I think this would be necessary to better adapt to specific situations...


      Best regards,

      Matthias Eller

        • Re: Hardware Monitoring Issues with Juniper Devices
          Víťa Tauer

          Hi Matthias, 


          thanks for feedback!


          Regarding 1), We're now treating Standby status as warning. We can reconsider it, the issue is tracked internally as 182051, we will keep you updated.


          Ad 2) - we're displaying the information that is passed to us directly by the device, in this case, the OID asked is If the device reports all of the fans as "Fan 1", there is probably nothing that could be done on Orion side, I think that this is configurable on the device.



          User defined thresholds aren't supported in current release of NPM. Reported status of entities (fans/temperatures..) is exactly what does the device tell us. The only exception in this are Cisco nodes, which reports the thresholds via SNMP as well. I can track this as a feature request for next version.


          Hope this helps


          -- Vita

          1 of 1 people found this helpful
            • Re: Hardware Monitoring Issues with Juniper Devices

              Thank you Youssarian,


              this helps, but I would like to add some thoughts.


              Regarding 1) I highly recommend reconsidering to treat Standby status in Juniper devices as up. Indeed standby does not mean that there is a problem with the component, but simply that it is not in use (failover ready). So the overall status of hardware monitoring in this case in my opinion should be "green" and under no circumstances generate "warning" alerts...


              As an example I state the temperature measure of the two Routing Engines in a Juniper EX4200 device. The standby routing engine is in a warning state, even if the temperature is normal. This makes no sense to me at all...


              User defined thresholds would be great!!


              Again many thanks