External Checks

replication-manage 2.0 can be used from proxies like HaProxy to check the status of the monitored servers.

We provide different checks for master and slave status.

https://{monitor-api-address}/api/clusters/{clusterName}/servers/{serverName}/master-status
https://{monitor-api-address}/api/clusters/{clusterName}/servers/{serverHost}/{serverPort}/master-status
http://{http-bind-address}:{http-port}/clusters/{clusterName}/servers/{serverHost}/{serverPort}/master-status

Return Code 200 if server id is a master

https://{monitor-api-address}/api/clusters/{clusterName}/servers/{serverName}/slave-status
https://{monitor-api-address}/api/clusters/{clusterName}/servers/{serverHost}/{serverPort}/slave-status
http://{http-bind-address}:{http-port}/clusters/{clusterName}/servers/{serverHost}/{serverPort}/slave-status

Return Code 200 if server id is a slave

{clusterName} is the name of the section name define in the config file default should be it if no multiple clusters are defines

{serverName} can be found via the client command topology

  • [x] Clusters in passive state return error code.
  • [x] Servers in maintenance, failed or suspect will return error code.
  • [x] Slaves with replication lags still return valid code
  • [x] Slaves with replication stopped return error code
  • [x] Master with READONLY return error code

Version 3.1.42 Addendum

The following endpoint change complements the legacy paths above, which remain documented for compatibility with existing scripts.

Version 3.1.42, PR #1747: reader-status for new HAProxy checks

Newly provisioned HAProxy externalcheck configurations use the following reader endpoint:

http://{http-bind-address}:{http-port}/clusters/{clusterName}/servers/{serverHost}/{serverPort}/reader-status

Return code 200 means that the server is an eligible reader. Existing proxies continue to use their generated slave-status script until they are reprovisioned.