11 Replies Latest reply on May 21, 2015 12:19 PM by txyang

    Auto-Detect/monitoring

    txyang

      Hello community.

      Question: Does solarwinds auto-detect new hardware via polling on machines (ex:HDD) and start to monitor those automatically? Or does it discover the new hardware via polling and then a manual intervention is needed to monitor the newly detected HW?

       

      Thanks!

        • Re: Auto-Detect/monitoring
          mrs.alterego

          Hi txyang - Orion will auto-detect, however it will require a manual action on your part to monitor the newly discovered/added devices. HTH!

           

          Rach

          • Re: Auto-Detect/monitoring
            bluefunelemental

            actually I do this today.

            Setup network sonar discoveries for your ip range or list of ip/hostnames with all of the options needed to import the devices.

            Set a recurring schedule but then don't run the discovery.

            GO into the DB manager, discoveryprofile table, then change the autoimport option to 1

            The next time it runs it will import in all interfaces and volumes even if they are new.

             

            now deleting out old orphan volumes and interfaces is something different.

              • Re: Auto-Detect/monitoring
                mrs.alterego

                Hey bluefunelemental - that's a really great work-around! I didn't know about that one, so thank you for sharing! :-)

                  • Re: Auto-Detect/monitoring
                    bluefunelemental

                    you flatter me but isn't that what this idea was all about?

                     

                    Discovery Profiles

                     

                    Right now I'm working on classification of discovery profiles and how to map them to nodes within our CMDB.

                    Classification - this one is somewhat easy as the choices available dictate the classes - generally switches(NCM,snmpv2 vs snmpv3) / servers (wmi, agent?) or ICMP for UCS,VMware, or VIP's then the community strings or credentials enumerate these out per system or site or region etc. It looks like we'll have say 36 discovery profiles per instance which is manageable.

                    -edit - I forgot to add additional polling engines. Assignment to which polling engine also defines what discovery profiles you setup.

                     

                    Identification - It's easy to push up this pick list of discovery profiles into our CMDB but when the wrong one is selected it's going to get messy real quick. I'm thinking of how to identify the correct discovery profile based on a custom property or better yet a combination of region, sitecode, vendor,  model, and machinetype but this is probably going to be a latter phase.

                     

                    Assignment - Here's where the magic happens. When the right discovery profile is selected then I'm working on appending the IP into the discovery profile plugin configuration so it now discovers and auto-adds everything from that node on it's normal schedule. I would rather not wait for the next scheduled discovery but it seems there is no easy way to kick off an existing discovery profile through API. I could query out the plugin config xml and feed that into a temp manual discovery profile but that sounds messy.

                     

                    Thoughts?

                  • Re: Auto-Detect/monitoring
                    txyang

                    Thanks for sharing!