Obviously you can't bend to every whim, but another of Network World's Best of 2008 got glowing reviews for it's ability to group results and report on the group's overall SLA (as opposed to each individual object's downtime). This sounds like an important feature (I have my own overall SLA's I'm trying to meet) and I think you guys have the core code in place to effect a crude version of an SLA model (using folders to group tests by SLA, apply dependency expressions for the group), but the trick is creating a formal interface to make this transparent. Also, tying a group of tests/devices together appropriately could be tricky depending upon the number of tests and each tests influence on the overall criticality.
The example given:
If one of five related servers suffers downtime, but the other four healthy servers continue to ensure application availability to the business community, [Competitor's product] accurately and correctly notes the server's downtime on its dashboard and in its monitoring reports. In addition, its SLA feature reports the overall availability of the shared-server application as "good."
For HM there may also be some modification of the dashboard presentation to accurately reflect this type of nesting/grouping.