Resource Measurement Facility (RMF) for z/OS is a critical component for performance monitoring and workload management in IBM z/OS environments. With the recent enhancements (APAR: OA67701) to RMF for z/OS versions 3.1 and 3.2, you can now leverage OpenMetrics to expose RMF Monitor III metrics in a standardized format. This capability opens new possibilities for integrating z/OS performance data with modern observability tools.

What is OpenMetrics

OpenMetrics is an open standard for transmitting metrics in a text-based format, widely adopted by modern monitoring solutions such as Prometheus, Telegraf, and VictoriaMetrics. By supporting OpenMetrics, RMF for z/OS enables seamless integration with these tools, allowing users to visualize and analyze z/OS performance data alongside metrics from other platforms. OpenMetrics delivers a robust monitoring framework tailored for mainframe environments when coupled with Monitor III for RMF.

Here’s an image of the data flow in an OpenMetrics ecosystem. The RMF DDS exposes Monitor III metrics via the OpenMetrics HTTP API. The third-party tools like Prometheus, Telegraf, and VictoriaMetrics polls the metrics at a regular interval by calling the OpenMetrics HTTP API. Lastly, Grafana leverages built-in support to these third-party plugins for metrics visualization.

• Enhanced Visualization: We ship production-ready starter Prometheus dashboards (also bundled into IBM RMF plugin) that provide immediate value for better performance management. 
• Industry Standardization: OpenMetrics is widely adopted by modern observability tools like Prometheus and Grafana, enabling z/OS RMF metrics to fit seamlessly into enterprise monitoring ecosystems.
• Unified Monitoring: You can now collect and store all performance data across your entire z/OS infrastructure into a single repository, enabling dashboards for a comprehensive view of operational performance.

Endpoints and Usage:

The RMF Distributed Data Server (DDS) exposes Monitor III metrics through the /metrics/m3 endpoint. In Prometheus, the endpoint is part of a "target" definition, which is just one part of this definition.

Also, the polling interval must match the Monitor III MINTIME value used for collecting accurate data.

The metrics supplied by the /metrics/m3 endpoint can be filtered using the groups parameter in the GPMOMC PARMLIB member which is shipped in the SYS1.SERBPWSV library.

You can retrieve sample requests for either all metrics or filters by specific groups such as CPCs or LPARs.

This example retrieves all available metrics in OpenMetrics format:

https://dds:8803/metrics/m3

This example retrieves specific metric groups in the OpenMetrics format, e.g. for CPCs or LPARs:

https://dds:8803/metrics/m3?groups=cpcs,lpars

Configuring OpenMetrics is easy and flexible, simply use the shipped YAML file to customize metric names and reload the settings dynamically without restarting the DDS.

The following image displays a YAML configuration snippet defining the CPC-related fields for OpenMetrics data collection.

For more information on configuring endpoints, refer to the OpenMetrics Endpoints documentation.

OpenMetrics transforms RMF from a specialized mainframe tool into a key player in enterprise observability strategies. It empowers teams to monitor z/OS alongside distributed systems using the same tooling, accelerating troubleshooting, and improving decision-making.

For more information on OpenMetrics, refer to the z/OS RMF IBM documentation.

Thank you for reading!