cluster alerts need instance grouping

We have a common use case in monitoring where an instance may exist on one or another member of a cluster (e.g., wireless AP on a pair of controllers). The cluster alerts mechanism does not account for this scenario (either counts devices or all instances, not status per instance).  The ServiceInsight feature may allow for this, if I can figure out how to make it work as needed, but it is a premium feature whose ecessive cost should not be incurred for dealing with this missing functionality.  Please add an option to cluster alerts to group instances across member devices for threshold evaluation.