I think I understand what this original post ment.... or not 🙂
HQ EE has this application type where you can define tree of resources depending on each others, but is lacking alerting functionality. Correlation adds it's own bits to this mess.
When you setup complex networks and application cluster which are all partly dependent on each others. When you monitor application A, it may not work because application B is down. This means that maybe there's nothing wrong with A(don't do control action).Worst situation becomes if you have 3 resources which all depends on each others. HQ can manage each resources, but only knows about one entity. It may be that when HQ is restarting these services, they never get change to get up.
Correlation would tell resource priority order for restarts. I've seen these things on some batch job management frameworks. It would be very sophisticated feature for HQ.