Q: How are fail-overs and alerts to system administrators supported?

A: Fail-overs are handled via heart beats directly from eZ Publish. The load balancer is connecting to eZ Publish and pulling the heart beat information from each node in the cluster. If the heart beat signal is not correct the load balancer simply excludes the node from the active cluster. Depending on the load balancer used notifications can be send to the administrator when a node is excluded from the cluster due to a failure.