Was doing some searches and actually ran across what may actually be the underlying issue here. When you created the Dependency monitors for the group, did you directly target the Availability and Performance aggregates (as one normally would?).
The reason i ask, is that based on https://blogs.technet.microsoft.com/volkancoskun/2014/10/19/how-to-fix-the-health-rollup-problem-in-distributed-applications-with-exchange-2013-health-sets/ and also a quick check in my own management group, the Exchange Service Component’s only unit monitor is an event monitor that has been set with it’s parent monitor as “Entity Health” directly.
As such, the dependencies are missing the only monitor that is actually generating health for these objects. The article proposes that you duplicate the monitoring, but for your purposes it may instead be easier to use the SCOM console to create a Dependency monitor that just rolls the top level Entity health of the service component up into Availability directly.