I am wondering if anyone has done any of the following, if so I would love to know how it worked out for you. Also if there is any code you can share, would appreciate it.
1) Has anyone configured MQ to be monitored via SNMP ?
2) Has anyone configured Zenoss to graph the metrics as well as be responsible for alerting ?
Also, anyone know a easy way to have the Agent Manager only send out alerts on state changes.
While we are not monitoring at the MQ layer, but we are processing ESB tracking messages and some container/queue level information using CA Wily.
You might be able to programmatically control how the alerts are sent (caching/comparing to the previous values). An example on the direct API use is at:
C:\Sonic\MQ7.6\samples\Management\runtimeAPI\javaJMX\MgmtCLI.java
You might want to try using the Zabbix appender for Netmon. Zabbix allows you to graph any metric you get into it and is also able to send alerts by e-mail and SMS. The appender is a bridge program between Sonic Event Monitor and Zabbix and will allow you to send any SonicMQ metric to Zabbix for graphing and alerting (for example: When the number of messages on queue x is over 5 for more than one minute, send me an e-mail. When it lasts for more than 2 minutes, send an SMS, etc).
http://www.psdn.com/library/thread.jspa?threadID=7011&tstart=15
Thanks, I will have a look. Do you know how hard it would be to port the bridge program for use with Zenoss as opposed to Zabbix.
If Zenoss has a publicly available network protocol it shouldn't be all too difficult. The Zabbix portion of the bridge is provided by a separate library (which is also open source) so the bridge program knows almost nothing about Zabbix.
You would only need to write such a library for Zenoss instead of Zabbix, which could be quite hard or very easy, depending on their network protocol and available libraries.