Better logging Log number of messages on heartbeat Log on connection from client Parse stack traces Use attrs https://attrs.readthedocs.io/en/stable/ Index types for fields What if ES goes down? Add tests for ConnectionHandler Error handling Garbage collecting finished handlers? Sample startup config Test running tcp server with tornado and connecting with syslog-ng Readme Upload to pypi Load testing Add project to a CI and add badges No indexing if there is no ES config