Monitoring queues

 

Very often the first sign of a problem with a queue that is being serviced is that the number of messages on the queue (CURDEPTH) is increasing. Of course, the increasing number of messages might not be a sign of a problem if it is expected at certain times of day or under certain workloads. However, if there is no known explanation for the increasing number of messages, it should be investigated.

You might have an application queue where there is a problem with the application, or a transmission queue where there is a problem with the channel.

This chapter discusses the various monitoring options that are available to determine the problem with a queue and the application servicing it. Monitoring channels discusses additional monitoring options that are available when the application that services the queue is a channel.

This chapter goes through the questions that you should ask, and the fields that you should look at in the output of various commands, to get the answer for your particular queue. In all the examples, the queue being examined is called Q1.

The questions detailed in this chapter are as follows:

• Does your application have the queue open?

• Are the messages on the queue available?

• Is your application getting messages off the queue?

• Can the application process messages fast enough?

• What about when the current depth is not increasing?