When I run haproxy with both TCP and HTTP health check on two backends it sends 503 Service Unavailable error and does not identify any of the backend server although they are up. Both the health checks work individually but when they both are setup together it sends error. Can these two health check work together in haproxy?
Technically, if the HTTP check fails a TCP connection then it will indicate the server is down. Therefore, you should not need both because theoretically the TCP check is built into the HTTP check.