Description:
MySQL Enterprise Monitor currently does not have any HA or failover mechanisms built in. We implement our monitoring systems in a redundant fashion, but MySQL Enterprise Monitor doesn't appear to have any sort of failover ability built into it.
How to repeat:
Stop monitor processes or shut down monitoring server. Agents then have nothing to talk to.
Suggested fix:
We may handle this internally by having a primary instance of the monitor which will be running tomcat and mysql. It will then have its mysql database replicated to a secondary server that will have the monitor installed but with tomcat turned off and the included mysql running as a slave. In the event of the failure of the primary, we'd promote the slave to master, start tomcat, and update DNS with the new IP that the agents will talk to. This is just a plan and has not been tested or implemented yet.
It would be nice to know if both tomcats could be running and if the mysql servers could be using master-master replication.
