Make it configurable how many times a node may fail to respond on the health check port before it is considered to be down.
The monitor now pings the health check ports of the Clustrix nodes. A response translates to RUNNING and a non-response to DOWN.