JMX notifications with EMS

Elias Ross genman at noderunner.net
Mon Aug 18 18:37:38 UTC 2014


Just a point. If I thought JMX notifications were useful, then I probably
would have patched EMS and added support to RHQ's JMX plugin.

I just thought the use case for notifications was not adequate.

For instance, you could have a JMX notification saying something went down
or some threshold changed, but then you can capture that anyway using a JMX
attribute, have full history of state, and not worry about that
notification getting lost.

So what might be useful is tracking a stream of text-based error messages,
doing filtering, and whatever. But then again, having a log file is
probably better as again you have a history of changes and it's visible to
more tools.

But then you run into the limitations of the event system in RHQ. So to me,
the 'ultimate solution' really is to fix the log-file based monitoring
using ElasticSearch as been proposed. You could easily just have the agents
'discover' all the log files of a system.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.fedorahosted.org/pipermail/rhq-devel/attachments/20140818/4d0d87b6/attachment.html>


More information about the rhq-devel mailing list