Our primary Spectrum server is on a ring network connected to our routing core via two VRRP interfaces. When the VRRP primary interface changes, Spectrum loses contact with the rest of the world for long enough to trigger a handful of loss of contact alarms. Needless to say, this is irritating.
I'd like to come up with a way to minimize the false alarm notifications. I've thought about adding a delay to every notification in SANM, but that would slow down every alarm for the rare occasion when it would help prevent false alarms and seems a bit much. I've thought about trying to kludge together something in Condition Correlation or Event Config, but the best idea I've got with that is having a custom event stop Alarm Notifier for a few minutes, which also seems a bit heavy handed. I've also thought about slowing down polling and re-tries for every model, but that also is a bit heavy handed.
How does everyone else handle situations like this?
Whenever is primary Spectum server is down it must switch to DR spectro server ....hope all devices must be monitored in DR Spectro server ..