I changed it to org.elasticsearch static one. The idea was to have each node a JMX aggregator of other nodes, but gave up on that. Here is the issue: https://github.com/elasticsearch/elasticsearch/issues/843.
As a side note, not a big fan of JMX, very heavy on system resources. The different stats/info/status APIs in elasticsearch gives you a lot of information.
Actually, those include, currently, much more info than JMX, so if you miss something, it can be easily added to JMX.
On Friday, April 8, 2011 at 5:27 PM, Paul Loy wrote:
we're in the last stages of productionising our solution. We are now facing a little issue with the monitoring of our app using JMX. Elasticsearch seems to be adding a random (enough) node id to the name of the ES MBean which changes. This means that we can't currently monitor ES.
Is there any way to override this behaviour such that the MBean is exposed with a static (or config derived) name?