[Not critical] Follow-up from "Return Default Probes in Duo Health Check if There Are Hybrid Models"

The following discussion from !199867 (merged) should be addressed:

• @manojmj started a discussion:

  @julie_huang sorry, I missed this during the first review, but it would be better if we move all of the new logic to inside this class. Not a blocker now, but we should definitely open an issue to follow-up. Let's encapsulate all of the logic in one place, that's better.