Fetch the required results only when when needed
    
    Fetches result for specific subject, test case and optionally scenario
    only when needed by a rule in policy.
    
    Only single result (`limit=1` instead of `limit=1000` in GET requests)
    is fetched at a time since it's very probable that it'll be the one
    needed. Next result needs to be fetched only if the result ID matches
    one in `ignore_result`.
    
    Since cache keys are now more specific (additionally test case name and
    scenario are used), fewer cache entries are invalidated when new result
    is added to ResultsDB.