5 Replies Latest reply on Aug 16, 2010 11:26 AM by chris.lapoint

    Feature Request - Understanding Device Dependancies

      This is a HUGE issue for us right now.  I have also seen other posts from other people regarding this issue.  SolarWinds proclaims to offer configuration of device dependancies in their sales promos, but when you get down to really using the product, it does not seem to offer a true understanding of device dependancies and keeping your email from being clogged from alerts on dependant devices when a core device goes seems impossible.  I have spoken with several technical support members from SolarWinds regarding this including case # 122831 and 122566 and have not received any method for truly handling/recognizing dependencies in NPM.

      For example...  if you have a core router --> distribution switch --> access switch --> multiple servers

      It would make sense that if the access switch went down then you would expect an alert on the access switch, but would not want hundreds of alerts on all of the servers attached to it.

      Likewise, if the distribution switch went down then you would expect an alert on the distribution switch, but would not want an alert on the access switch or any of the servers that rely/depend on the distribution switch.

      NPM 9.5 SP5

      Thank you for your assistance/consideration.

        • Re: Feature Request - Understanding Device Dependancies
          denny.lecompte

          This is a high priority for us.  One question:  For a switch, figuring out the device dependencies is something that can be done somewhat automatically by analyzing the topology.  But what about when an edge router goes down?  We're looking for a way to allow users to define dependencies in a simple UI, but that type of definition would be manual.

          Once we have both manual and topology-based dependencies defined, we'll set status based on those dependencies:  If a parent device goes down, the child devices would be marked as unreachable rather than down.

          Is that what you're looking for?

            • Re: Feature Request - Understanding Device Dependancies

              That sounds like it may meet our needs, but honestly, it is something we will need to use before I know for sure.  At the very least, it sounds like an enormous improvement over the condition/suppression abilities that exist now.

              I would expect that an edge router could be handled the same way as a switch.  Orion should be able to realize either by mapping the topology of the network (preferred) or by allowing input of dependency/logical topological info which devices are connected to which and how.  Once it has that info, it should be able to automatically suppress alerts for dependant devices down the line from a device that has gone down.  For example...  if you have remote sites A and B connected back at a main data center to a core router, and each site has its own router, distribution switch, access switch, and any number of end devices (servers, workstations, printers, etc.).  Then, if at site A, the distribution switch goes down (or otherwise becomes unreachable, Orion will recognize this topology (and the correllated dependancies) and alert only on the distribution switch and NOT the access switch or any of the monitored devices that are connected to the access switch.  Likewise, if the access switch at site B goes down (same time or not) Orion will alert on that access switch, but NOT any of the devices connected to it that are being monitored.  Similarly, if the core router back at the Data Center goes down, then Orion will alert on the Core Router and not any of the other devices that are dependant on it, for example, site A and B.  However, the solution must also be able to recognize redundant links in such a way that if you have a ring or mesh topology and a single route to a remote site goes down, but other routes are still available, then an alert can be sent on that link being down, but no other alerts are sent.

              I hope that clears things up a bit more?  Please let me know if you need any other clarification as to what we are looking for with regards to this matter.

              Do you have any approximate dates on when these features may become available?

              We appreciate your assistance in getting this feature out ASAP as it is really a HUGE deal for us.

              Sincerely,

              -Steve

                • Re: Feature Request - Understanding Device Dependancies
                  denny.lecompte

                  Orion should be able to realize either by mapping the topology of the network (preferred) or by allowing input of dependency/logical topological info which devices are connected to which and how. 

                  In the case of switches, it's a lot more obvious that if the switch is down, then whatever is plugged into it can't be reached.  Once you start dealing with Layer 3 dependencies, there are often multiple routes to a given device, so figuring it out automatically is not always possible and calling a device "unreachable" when it's actually Down is unacceptable.  Having those dependencies input manually is more reliable.

                  And it is a big priority for us, although it's a major effort.  It should appear in 2010.