[
https://issues.apache.org/jira/browse/AMQ-4237?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14225302#comment-14225302
]
Sam hendley commented on AMQ-4237:
----------------------------------
There is a downside to this approach. I believe it is now impossible to run a
JMX query that selects the Queue/Topic/Broker without getting every single bean
"below it" in the hierarchy. This makes monitoring the broker via JMX nearly
impossible when there are many consumers. This also leads to many error and
exceptions as we try to collect MBean information for short lived consumers
which we can't avoid selecting in our queries.
For reference here are the JMX queries I use:
org.apache.activemq:*,destinationType=Topic
org.apache.activemq:*,destinationType=Queue
On 5.6.0 polling this took around 500ms for our few hundred queues+topics. Now
the query takes around 10-15 seconds since it also collects all of the
consumers. That being said I really like being able to dig into the per queue
consumers with JConsole when debugging but it really makes monitoring a pain,
so much so that I have given up collecting broker stats on most of our systems.
I have searched far and wide and haven't been able to find any ObjectName query
that allows exclusions, maybe there's something else that can be done I don't
know about?
> JMX ObjectNames do not follow JMX Best practices to use a Hierarchical format
> -----------------------------------------------------------------------------
>
> Key: AMQ-4237
> URL: https://issues.apache.org/jira/browse/AMQ-4237
> Project: ActiveMQ
> Issue Type: Improvement
> Components: Broker
> Reporter: Rob Davies
> Assignee: Rob Davies
> Fix For: 5.8.0
>
>
> The Current JMX ObjectNames result in a disjointed view from JConsole of the
> managed objects and services of the ActiveMQ Broker. By following best
> practices, it will be easier to find and monitor all the endpoints associated
> with a Destination - for example.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)