Previously known as MQSeries®, IBM® MQ is an enterprise grade messaging system used to transfer data between application services through queuing middleware. It supports several modes of operation including point-to-point, publish or subscribe and file transfer.
Applications can also publish messages to many subscribers over multicast. IBM MQ monitoring requires the ability to identify any IBM MQ instances, collect performance information from the platform and data about the queues and queue managers.
With IBM MQ, if a message can’t be delivered immediately, it’s secured in a queue, where it waits until delivery is assured. IBM MQ acts as a shock absorber between applications in asynchronous messaging environments. If there’s network or application disruption, it holds messages in special queues until everything is running.
*Visualization of publishing and subscribing architecture courtesy of IBM MQ developer website.
When the Instana agent is deployed onto a host that contains IBM MQ, the Instana IBM MQ sensor is automatically configured and attached to IBM MQ. Instana then immediately begins monitoring IBM MQ performance metrics to ensure that smooth messaging is maintained. Instana gathers IBM MQ performance metrics for clusters, listeners, topics, queue managers, queues and channels all with 1 second granularity so you don't miss a spike that indicates a potential performance problem.
Instana maintains a curated KnowledgeBase of health signatures for all monitored platforms, including IBM MQ. The KnowledgeBase is continuously evaluated against the incoming metrics and are used to alert on issues or incidents depending on user impact and severity. Built-in events trigger issues or incidents based on failing health signatures on entities, and custom events trigger issues or incidents based on defined thresholds of any entity’s individual metrics.
Event
- Queue Manager number of connections
- Queue Manager status
Description
- Check whether there are no connections currently on the queue manager
- Check whether queue manager is healthy and not in a quiesced state
Metric
- Connection count (connectionCount)
- Queue manager status (status)
Event
- Queue oldest message
- Queue depth diff
Description
- Check whether there are messages on the queue older than the threshold value
- Check whether queue depth is approaching max queue depth
Metric
- Oldest message on queue (oldestMessage)
- Queue depth (queueDepth) and max queue depth (maxQueueDepth)
Event
- Channel status
Description
Check whether the channel is in a healthy state
Metric
- Channel status (channelStatus)