Monitoring IBM WebSphere MQ Servers

eG Enterprise prescribes a specialized WebSphere MQ monitoring model (see Figure 5) that monitors the performance of the core WebSphere MQ components such as the queue manager, the message queues, and the channels, proactively identifies slowdowns, and alerts administrators to the anomaly before any loss of communication occurs.

Figure 5 : Layer model of a WebSphere MQ server

Each layer of Figure 5 is mapped to a wide range of tests, which execute on the WebSphere MQ server and collect performance statistics that reveal the following:

  • Is the WebSphere MQ server available? Are clients able to connect to the server?
  • Are any of the queue managers unavailable?
  • Is there a delay in connecting to any queue manager?
  • Have too many messages been enqueued?
  • Are the queues processing messages too slowly?
  • Is the local queue full?
  • Are clients able to connect to the channels?
  • How is the load on the channels?
  • When was the last time the channel transmitted a message? Has it been too long since then?

The sections that follow elaborate on the tests mapped to the top 4 layers of Figure 5. For details on the other layers, please refer to the Unix and Windows Servers monitoring model.