Interface Alerts triggering off when Switch going into a Down Status

rjnicholson

I am getting alerts firing out off my 2 different advanced alerts for Node and Interfaces when the Node drops.

More detail.  I have custom properties I have set in the database for our Nodes/Devices to be grouped by Department.  In this case devices listed with DataComm in the field since I am dealing with a Switch in this case.

I have 1 rule for Node Down

Trigger Alert when all of the following apply

     Department is equal to DataComm

     Node Status is equal to Down

I have another rule for interfaces

Trigger Alert when all of the following apply

     Department is equal to DataComm

     Interface Status is equal to Down

Today I had a switch fail and it triggered alerts on all active interfaces and then on the actual down node.  I DON'T have dependencies built off of these switches, but was thinking (actually from the 10.1 release notes I'm 100% sure) that it should be an implicit dependency since the interfaces are reliant on the switch being up to be in an up status as well.

Just checking but is this just a part of the polling issue on dependencies implicit/explicit that has been fixed on 10.3 release which will roll to next week.  Currently on 10.2, or are my advanced alerts a bit to general and need some tweaking.

.

mrxinu

Your understanding of an interface's status going to 'Unknown' when the node where it resides goes down is consistent with what I know.  This bit from the 10.3 release notes seems to speak to that, though.  Maybe someone from SWI can chime in.

"Defined dependencies are now properly respected in the event that a dependent object is polled before its parent object, ensuring that alerts are triggered in accordance with existing dependencies."