It is often desirable to monitor and display services that are provided by a high availability cluster.
In normal operation all nodes would be in the OK state. If the service fails-over to another node, it is vital to know this has happened as you are now at a single point of failure.
The objective is to monitor the portable parts of the "service" on a node and also include the hosts as nodes below the service.
For this example we will use a SQl cluster.
To do this we need to create the "virtual service node" as a partial duplicate of the primary host.
To do this, add a new node in the node manager, (make it ping only and allow duplicates) and use the floating or "VIP" IP address for the service.
1. Select a point in your node tree that best describes how it is connected to the network. Right-click on that point.
2. Once the node has been added, add a new IP SERVICE Test for "openPort" and add the port number (normally 1433)
Depending on what the service consists of, other service tests can be added to prove it is operating or not.
3. On the current primary host, open ping group and ensure both IP addresses are ticked on for polling (physical and service IP address).
4. Now drag the the 2 hosts to make them children of the new service node. (change the ICON on the service node)
5. Add the group of nodes to a view and turn on the draw heirachy option for the view. Drag the nodes on the view to represent the Hosts under the service.
If the system fails over, the primary node will go critical as the IP service port test will fail, but if all is good, the service node will remain green and the failed hosts will be warning or critical.