It appears that we have not (yet) assembled a breakdown for each alert classification. I'm on it. It will take some time for me to collate and annotate this.
In the meantime, attempting to address your specific question, if you are asking about the CAPWAP delay time alerts, our logic is
if the round-trip echo time exceeds ~60ms, then send a minor alarm. If it exceeds ~123ms, then send a major alarm. A critical alert is one that we deem needs urgent operator intervention to remedy (i.e. the CAPWAP connection between HiveManager and the managed device was broken).