i hear you, all these points make sense. but we have run into issues where a node or more commonly an interface was up/normal and then changed into unknown state which obviously suggests that theres an issue & there isnt any way currently for nnmi to trigger an alert on this....
The following is a workaround for "Unknown" status which doesn't send alerts.
Make sure all nodes are a part of the "Important Nodes" group. Any member of the Important Nodes group will not go into Unknown status if unavailable/down. Instead it will go into Minor/Major/Warning/Critical status.
In the Additional Filters tab in the Important Nodes group, create a rule to include any node you wish that will no longer go into unknown status.
Once those nodes are a part of the Important Nodes group, you will always recieve the critical alerts attributed to the group.
I've never been able to figure out why HP designed NNM this way.