Dynamic workload management, which is performed by the Workload Manager (WLM), is an essential part of the z/OS operating system where multiple workloads run simultaneously.
Tip! Read this white paper: Managing Workload Manager Goals and Performance
This brief video demonstrates IntelliMagic Vision for z/OS reporting for WLM.
Workloads are categorized into distinct service classes, for which a ‘goal’ and ‘importance’ is defined. The goal describes what performance a service class should get, and the importance defines how important it is that goals are met for this service class, relative to the other service classes. WLM uses this to make sure the most important workloads get priority at times when resources are scarce.
The performance index (PI) is the metric that measures whether WLM goals are met, by computing how high the response time was relative to the goal: PI = Measured Transaction Response Time / Response Time Goal. A value greater than 1 means the response time for the service class did not meet the goal.
Intelligent Ratings Highlight Potential Issues
IntelliMagic Vision for z/OS makes a rating for how well goals were met, based on the height of the PI in combination with the importance. This gives clear insight into how well WLM manages to achieve its goals, and in case of missed goals, how severe problems are. These ratings are shown in the WLM dashboard.
Understanding Internal Components with Deep Insights
When the dashboard shows yellow or red bubbles, it is important to be able to find out why the response time goals were not met. The “using” and “delay” components of the response time are shown by IntelliMagic Vision at various levels, such as per LPAR or per service class. This provides a very deep insight in whether CPUs are busy doing actual work, or whether there is any contention on the processors or on the DASD.
For instance, when the Using DASD component is significant, it is very worthwhile to investigate using IntelliMagic Vision drill-downs if I/O tuning or a (partial) storage system upgrade can make a difference. These are likely much less expensive than a processor upgrade.
Monitoring TCP/IP Performance for z/OS with SMF 119 Metrics
Gain visibility into the mainframe TCP/IP communications infrastructure performance with SMF 119 metrics. Numerous activity and utilization metrics are reported.
Avoid Inefficiencies with Store Into Instruction Stream (SIIS) Assessments
Programs that store data into the instruction stream introduce major inefficiencies into CPU performance. Updating SIIS programs can result in significant reduction of CPU utilization.
Reduce Latency and Lower Demand as You Create the Capacity Plan
Mainframe capacity planning is an important budgetary input. If you ensure your performance is in order or have mentioned some efficiency recommendations, you will help keep your organization lean and productive.