Hi,
We're trying to use Glassfish self management capabilities to implement infrastructure monitoring at the "middleware" layer. The objective is to integrate it with a CA Unicenter operational environment.
We would be interested in experiences in application server monitoring.
I can see it as two sided: Application "heartbeat" and statistics trough URL polling or jmx based collection; Application server data collection through JMX/self-management.
We have a tentative notification mbean. It is surprising no example or default mbeans exists. The mbean writes lines to log files parsed by Unicenter or starts unicenter commands.
What would be best practices in application server monitoring. I see:
- JVM usage (using percentage would be peferred)
- CPU
- Application status
- Ressources (connectors,queue)...
As basic categories to be filled from observing jconsole. Does anybody have experience with self-management or other monitoring tools.
In addition, It is not clear to us what parameters are available to a notification mbean t access rule info (level,...) or the actual event information.
Olivier.
[Message sent by forum member 'duboiol' (duboiol)]
http://forums.java.net/jive/thread.jspa?messageID=321793