DX Application Performance Management

Expand all | Collapse all

Introscope Agent for WebSphere Message Queue and Broker

  • 1.  Introscope Agent for WebSphere Message Queue and Broker

    Posted 09-04-2013 09:06 AM
    I am looking for feedback on what anyone/everyone out there has done in their environment:

    1. for WMB and WMQ monitoring, deployed introscope agent on a separate server to monitor them remotely
    2. for WMB and WMB monitoring, deployed introscope agent on the same server where WMB/WMQ is running

    Please let me know when and why not?

    Thanks in advance.
    Manish


  • 2.  RE: Introscope Agent for WebSphere Message Queue and Broker

    Posted 09-06-2013 02:09 PM
    No one has any feedback on this or any experiences or best practices???


  • 3.  Re: Introscope Agent for WebSphere Message Queue and Broker

     
    Posted 07-29-2014 11:01 AM

    My first consideration is the size / scope of the WMQ/MB environment to be monitored. How many queue managers, brokers, etc across how any hosts? It's important to understand how many objects you will be monitoring from a single APM Agent and in turn how many metrics that single Agent will report into the APM Enterprise Manager since that will also vary depending on Minimum/Recommended/Full configuration.  (There are clamps built-in to avoid flooding but you want to avoid any constraint there).  Likely you will then have a better idea how many potential Agents will be deployed (i.e. 10k metrics - one agent is fine; 60k metrics - you hit the clamp, so split it into two agents at least, maybe more).

     

    In your organization, is it easier to manage multiple agents all co-hosted (say with a Collector EM)? If so, be sure that host has adequate CPU, memory so the Agents don't impact the EM processing.  If it's easier for your organization to keep Agents hosted on the box/system they are monitoring, then that has the benefit of Agents distributed across the environment (maybe 1-2 per host depending on configuration) vs loading up multiple Agents on a single host.


    My second consideration is security. Running from the MQ/MB host or an MQ Admin host means the 3rd party libraries should all be available to add to the Agent configuration as well as that host should include an account with appropriate permissions to run the Agent(MQMonitor/MBMonitor).