Performance events

Performance events are notifications that a resource has reached a threshold condition. For example, a queue depth limit has been reached.

Performance events relate to conditions that can affect the performance of applications that use a specified queue. They are not generated for the event queues themselves.

The event type is returned in the command identifier field in the message data.

If a queue manager tries to put a queue manager event or performance event message on an event queue and an error that would typically create an event is detected, another event is not created and no action is taken.

MQGET and MQPUT calls within a unit of work can generate performance events regardless of whether the unit of work is committed or backed out.

The event messages for performance events are put on the SYSTEM.ADMIN.PERFM.EVENT queue.

There are two types of performance event:

Queue depth events

Queue depth events relate to the number of messages on a queue; that is, how full or empty the queue is. These events are supported for shared queues. The queue depth event messages can contain the following event data:

• Queue Depth High
• Queue Depth Low
• Queue Full

Queue service interval events

Queue service interval events relate to whether messages are processed within a user-specified time interval. These events are not supported for shared queues.

IBM MQ for z/OS supports queue depth events for QSGDISP (SHARED) queues, but not service interval events. Queue manager and channel events remain unaffected by shared queues. The queue service event messages can contain the following event data:

• Queue Service Interval High
• Queue Service Interval OK