Would be pretty easy as you already have service checks that do this, but instead of just validating customer sql against their expression, you could add a timing for connection time, & response to query time. Let users specify their own SQL command to check, and then which servers / systems / cluster to launch it against.
