Introduction to Automated Change Detection for z/OS Performance Management

Role of Automated Change Detection

Today’s z/OS environments tend to be very complex as they support numerous business critical workloads with high availability requirements. Despite the fact that the mainframe generates an extensive set of SMF and RMF metrics that far exceeds other platforms, it can be a challenge to effectively leverage that data to proactively manage your ever-changing environment while moving at the fast pace of business today.

To help enhance the value you can derive from that SMF data, IntelliMagic Vision provides two complementary types of automated anomaly detection.

• Health insights views, which leverage best practices thresholds to detect when underlying infrastructure is at risk of not delivering responsive service, and
• Change detection views, which detect when significant changes occur in workloads or key metrics

Use Cases for Automated Change Detection

The videos in this series present 8 sample use cases showing how Change Detection can proactively enhance availability and performance in z/OS environments.

In addition to the anomalies identified through the Change Detection capabilities, these examples also illustrate the GUI-based, interactive reporting provided by IntelliMagic Vision that leverages dynamic navigation and context-sensitive drilldowns, which facilitates rapid and focused analysis to investigate any identified anomalies.

1. CPU by System
2. CPU by WLM Importance Level
3. CICS CPU per Transaction
4. Db2 Getpage Rate
5. Coupling Facility Request Rate
6. Batch Job Elapsed Time
7. CICS Max Tasks
8. Health Insights and Change Detection

Review of Considerations Relative to Change Detection

Here are some key findings and considerations arising from these Change Detection use cases:

• Deviations are often driven by application changes, and thus it is helpful to have tooling to easily identify drivers (e.g., Db2 plan or auth ID, CICS tran, address space, service class)
• Sustained deviations are of particular interest, in which case Change Detection provides the opportunity to get an early jump on analyzing and potentially remediating any negative impact
• Dynamic navigation is particularly helpful to enable changing the direction of analysis midstream, based on findings during the course of investigating
• Transaction deviations in Dev/Test may be of interest in some scenarios and not in others, highlighting the need to separate transitory issues from sustained changes that are destined to negatively impact Production if not remediated before that migration occurs
• Batch Jobs are likely to generate many deviations, so it is important to be able to focus that analysis on jobs on the critical path of key business cycles