Dependency Alerting Failure

coreys

In the dependency, the Parent is a group which contains 1 node (WAN router). The Child is a group with sub-groups containing every node upstream of the WAN router. There is an e-mail alert that triggers when the status of any node is "down".

Recently, the e-mail alert for the WAN router "down" triggered, but the e-mail alert for all of nodes in the nested Child group also triggered.  Is there a better way to make sure the e-mail alerts for Child node "down" do not trigger in this dependency?

designerfx

2 things which may affect things slightly. 3rd I realized as I was typing this up!

1: You don't need a group as a parent dependency if it's only a single node/interface/etc being the parent. You're still going to be naming the dependency anyway, so this may be a moot point. If you have multiple devices in the parent group (Such as redundancy), then you want a group for the parent.

2: When the parent goes down, it can take another polling cycle or two (depending on how the GROUP is configured) for the group itself to then be unmanaged, especially if you are editing it or changing something at the same time. This has been my personal experience, but it may not be de facto for NPM. You may get some alerts prior to this point, but you shouldn't be getting an alert for every single node. Did you get one for every single node or just a few before they were unmanaged?

3: what is your status configuration for the parent GROUP? Is it mixed shows warning (see advanced settings)? Because if so, that won't unmanage the child nodes immediately.

Also, wouldn't you want separate custom properties to define the parent nodes alerts vs the child nodes?