I've run into a weird issue today, which I'd like to attribute to a fluke and resolve. I've set Data Collection to notify on certain groups of nodes for sysUptime changes (reboots basically). In the past this has worked perfectly. A node gets polled for sysUptime and if it returns a lesser value than the previously stored value, the system (or SNMP Agent) has restarted.
During the week a request was put forth to create a maintenance window for a few devices. These devices did not show up in the topology so I notified the requester that they should be all set and closed out the ticket. These devices are for testing and usually are not monitored unless specifically requested.
Low and behold, I get an email from the requester stating he received an alert regarding the device rebooting. At first I didn't believe it and assumed the notification was from some other monitoring tool. I checked the Alarm Browser, and there it was, two alerts from two devices that don't exist in the topology.
I did an ovtopodump and confirmed the devices do not exist. The only place that they do exist is in the Event Configuration. Now it has been my experience that Data Collection will only work on nodes that exist in the topology. So why would I still receive an alert for a device that doesn't exist in the topology?
Any ideas? 10pts to someone who can explain why this happened.