The Alerts feature has a number of alerts that are specific to Agent Watcher - alert when a monitored service is down, alert when a monitored file is missing, that sort of thing. I haven't found anything similar for Agent Health.
I was thinking I could just run an Affected Computers report against the vulnerability group I've created for Agent Health, but if AH is doing its job that will always report 0 machines, since they'll in theory be repaired immediately.
My ultimate goal is to wait a few months and then generate metrics on which machines are having the most problems with Agent Health so I can try to more proactively troubleshoot. I could also look at which problems are appearing most frequently. I can't run either of those reports at present because I just don't have any data to work with.