With the job node metrics you can monitor a master node's connection with its job nodes. The object name is available on all nodes, but it only makes sence to monitor this on the master node. There are two sets of attributes to monitor. One set provides an aggregated view of all the job nodes while the other provides individual attributes for each job node.
Communication errors are only reported if the server uses the General queue. If the Hight throughput queue is used, the master node never contacts the job nodes on its own initiative and therefore there is no support for monitoring the job nodes through JMX in the current version of the server.
Apart from communication error attributes, the set of individual attributes also includes information about the host and port of a given job node. This information is not meant for monitoring as such, but is included for convenience, as when an error does arise identification of the job node involved is usually necessary.
- Communication failed
- Seconds since communication failed
- Max seconds since communication failed
- Number of job nodes
- Number of failed job nodes