Dynatrace

Dynatrace gives IBM MQ teams full-stack observability: host CPU, disk latency, queue manager process health, queue depth, and channel state on one platform with Davis AI highlighting probable root cause. OneAgent installs once per host or node and auto-discovers processes including amqzxma0 and related agents, while the IBM MQ extension (or technology module per your Dynatrace release) issues PCF or API reads against queue managers you authorize. Beginners enable the extension, see depth tiles, and page MQ admins without checking whether depth rose because publishers accelerated or consumers stopped—Davis may already link a stopped Kubernetes deployment to a growing ORDERS queue. This tutorial covers OneAgent placement for distributed and container MQ, extension configuration fields explained, built-in MQ dashboards, problem detection versus custom metrics, integration with Prometheus metric ingestion if used, security for monitoring credentials, and how Dynatrace fits beside MQ Console and open-source Grafana stacks.

Architecture: OneAgent Plus MQ Extension

  1. OneAgent on MQ host collects infrastructure metrics and process availability.
  2. MQ extension connects to queue manager via client channel with monitoring user.
  3. Metrics flow to Dynatrace Grail or classic metric ingestion (version-dependent).
  4. Davis baselines depth, channel state, and host resources; opens problems on deviation.
  5. Operators drill from problem to queue entity to host disk to consuming service trace.
Configuration concepts (extension setup)
SettingPurposeBeginner tip
Queue manager nameIdentifies which QM to pollMust match exactly as in strmqm display name
Connection name / host / portPoints to listener for client connectionTest with runmqsc or client sample from same host as OneAgent
Channel nameSVRCONN used by collectorDedicated MONITOR.SVRCONN, not production app channel
CredentialsAuthentication to QMVault-stored; DISPLAY-only authorities
Queue / channel filtersLimits polled objectsReduces noise and extension load on large estates

Metrics and Dashboards

Dynatrace provides out-of-the-box IBM MQ views when the extension is active: queue depth heatmaps, channel status tables, queue manager availability. Custom dashboards can combine MQ depth with host CPU—critical because depth alone does not prove MQ is the bottleneck. A queue manager at ninety percent CPU may throttle puts; Dynatrace shows CPU saturation on the same timeline as rising depth. Message put and get rates, when available, help distinguish publisher flood from consumer outage: puts up and gets flat implies missing or stuck consumer.

Davis Problems for MQ Incidents

Davis learns seasonal patterns. A Friday payroll batch may normally raise depth on PAY.QUEUE; Davis expects that shape. A Wednesday afternoon spike triggers a problem with likely cause candidates ranked—disk full on /var/mqm, channel RETRY on partner link, or application error rate up on a connected service. Operators validate ranked causes rather than trusting AI blindly: false positives happen after application architecture changes until baselines relearn. Document maintenance windows in Dynatrace so planned channel bounces do not open problems.

Deployment on Kubernetes and VMs

  • VM: OneAgent installed on OS; extension config points to local or remote QM listener.
  • Kubernetes: Dynatrace Operator injects OneAgent into MQ pods or node monitoring per pattern.
  • Multi-instance QM: monitor each instance and shared storage health—failover events should appear clearly.
  • Remote z/OS: often via mid-tier collector; confirm supported topology in Dynatrace MQ docs.
text
1
2
3
4
5
6
Validation steps after enabling MQ monitoring: 1. Dynatrace → Technologies → IBM MQ → confirm queue managers listed 2. Compare one queue depth tile to: DISPLAY QLOCAL('TEST') CURDEPTH 3. Stop a test consumer; confirm depth problem opens within baseline window 4. Restore consumer; confirm problem closes and depth drains 5. Add custom metric ingestion only if architecture requires Prometheus federation

Custom Metrics and OpenTelemetry

Some teams export Prometheus metrics from an MQ exporter into Dynatrace via Prometheus metric ingestion API. Unify alerting carefully—do not alert on depth in both Davis and Grafana Alertmanager without coordination. Prefer Dynatrace as single pane when enterprise standard, or keep Prometheus for cloud-native tiers and Dynatrace for mainframe MQ gateways only.

Security and Compliance

Monitoring users need read-only MQ authority. Audit who can view Dynatrace MQ dashboards containing queue names that imply business functions (payments, fraud, healthcare). Role-based access in Dynatrace separates NOC view from middleware admin view. TLS cipher suites on monitoring channels should match corporate policy. Log Dynatrace configuration changes in change management like any production monitoring change.

Troubleshooting Missing MQ Data

  1. Extension active? Check technology status in Dynatrace UI.
  2. Can host telnet or openssl s_client to listener port from OneAgent host?
  3. CHLAUTH and CONNAUTH allow monitoring channel and user?
  4. Queue manager running and name spelled correctly?
  5. Firewall between collector and QM after datacenter move?
  6. Extension version supports your MQ fix pack?

Explainer: Smart Smoke Detector

Dynatrace is a smoke detector that also guesses whether the kitchen fire started on the stove or in the toaster—operators still grab the extinguisher (runmqsc) but run to the right appliance first.

Explain Like I'm Five: Dynatrace

Dynatrace is a robot that watches your LEGO bins (queues) and the kids (programs) playing near them, and rings a bell when a bin fills up weirdly fast while also pointing at which kid stopped taking pieces out.

Practice Exercises

Exercise 1

List five Dynatrace entities you expect in a problem about growing queue depth.

Exercise 2

Write a monitoring channel naming standard and authority profile for read-only collection.

Exercise 3

When would you keep Prometheus alongside Dynatrace for the same queue manager?

Frequently Asked Questions

Frequently Asked Questions

Test Your Knowledge

Test Your Knowledge

1. Dynatrace OneAgent primarily provides:

  • Auto-instrumented host and process metrics
  • MQSC scripting
  • COBOL compilation
  • JCL submission

2. Davis AI helps MQ operators by:

  • Correlating anomalies across stack
  • DEFINING channels
  • Clearing DLQ automatically
  • Disabling TLS

3. MQ extension in Dynatrace typically needs:

  • QM connection and credentials
  • Only FTP
  • CICS region name
  • JES job class

4. High queue depth in Dynatrace should trigger:

  • Runbook with MQSC and app checks
  • Immediate DELETE QLOCAL
  • Removal of CHLAUTH
  • Disable logging
Published
Read time21 min
AuthorMainframeMaster
Verified: Dynatrace IBM MQ extension docs, IBM MQ 9.3