Does this imply that we have to constantly poll through this interface to determine that bad things are happening in the Glassfish container? My response is off-topic with respect to directly commenting on the REST API. But, I would like to suggest that it would be more useful to be able to define alarm conditions (resource depletion, non-responsive applications, etc) and then have something on the container side PUSH events/alarms to the monitoring application through a standard messaging API (e.g. JMS) instead of having to constantly poll to determine that everything is ok 99% of the time.
[Message sent by forum member 'ametzger' (ametzger)]
http://forums.java.net/jive/thread.jspa?messageID=286901