5 Replies Latest reply on Aug 15, 2008 2:44 PM by wdmail

    Dependencies setup - suggestion/question

      I'm in the process of setting up dependencies for each device and also routers for remote sites that the devices will be dependent upon.


      First....a suggestion - When setting up a device or group dependency, I'd like the system to automatically assume I'm interested in the monitors currently assigned to the device or group I'm working with. Right now, the filter assumes whatever the previous filter was set as.

      There is already an easy option to use a "previous" filter in the event that function is needed. (ie. Last filter used or previously saved filter.)


      Now the question - Is there an easier way of setting up the dependencies? Currently, I'm selecting the device/group and using the "add dependency" drop down for each device/group.


      Another question, though I'll have tested it shortly. How is the system going to react when I create a group of devices by geographic location, set the router monitor as a dependency for the group, where each device in the group already has it's own dependency (ping)? 

       My first assumption is to make sure the timing/warnings are set at a more frequent interval for the router than the device ping...and the ping at a more frequent interval than the other monitors.


      FWIW - I really like the way the dependencies show up on the NOC view. Even though we know the other monitors are technically "down", it doesn't generate more confusion (and red lines) for those looking at the screen and making phone calls to the on-call.

        • Re: Dependencies setup - suggestion/question

          Been thinking about this a bit more....

          Maybe make it so we can set a particular monitor as a dependency of many in reverse. In essence, "add existing monitors/devices" to the dependency monitor.

          Rather than adding the dependency to the group, you add the "group/device/monitor" to the dependency. As is done with alerts. We can select an alert, then add whatever we want to it.

            • Re: Dependencies setup - suggestion/question

              Hello Wdmail,

               Essentially you want to organize your Monitors and Devices based on the trip taken to get to these.  For example, let us say you are monitoring the Bandwidth and the Drive Space on a Mail server in a specific office and these tests depend on a few appliances such as a Gateway and a Switch.  In order to configure your dependencies accordingly, you would first create a group for this office, then a subgroup for the Mail Server Device.  You would then create a Ping Monitor to the Gateway and a Ping Monitor to the Switch (if managed).  As Monitors cannot exist outside of Devices, you would need to put these into Devices.  Add these Devices and the Mail server group directly into the Office group.  Then add the Mail Server device to the mail server group and within the Device, in addition to the bandwidth and drive space Monitors, add a Ping Monitor and make it a dependency for the Device.  Now add the Ping Monitor to the Switch as a dependency to the mail server group and add the Gateway Ping Monitor as a Dependency to the Office group.  As a result, if a Dependency fails, anything in that group will be supressed from Alerting.  Here's a screenshot that should speak better than words.


              Hope this helps organise your Dependencies as needed.



              1 of 1 people found this helpful
                • Re: Dependencies setup - suggestion/question

                  Thanks for the reply. Let's see if I have this right....(the picture is a welcome addition :~) )


                  A dependency will effect the sub-groups below it and the one it's contained within.

                  In application, if the dependency monitor triggers within a group, all groups nested within, including the group that contains the dependency, will not trigger other alerts.

                  In your scenario....

                  If the gateway ping fails, only the gateway ping alert is triggered. All others are suppressed.

                  If the switch ping fails, the monitors of the mail server are suppressed.

                  At the mail device level, the ping dictates whether the other monitors alert. 


                  This being my understanding, if correct, it sure would be nice to be able to set a dependency on a smart group.