I'd like to sit a bare-metal cluster in front of an F5 LB and configure some healthchecks for the kube-api / master nodes using the /healthz endpoint. On a successful, we get the HTTP 200 response with the body 'ok', but can someone provide what sort of response we would get in terms of failures? Also, does the HTTP response stick to 200 on a failure, and simply provide a different response? My concern here is if a failed check returns HTTP 200 OK, but 'bad' as a response in the body, this would not work on an F5 as a result of being case insensitive.