Using contexts

Edit online

A context refers to the specific circumstances in which data analysis and predictions take place. For instance, the z/OS workloads, network traffic, JVM garbage statistics, and CICS region transactional activity are all contexts that shape how the machine learning algorithms within OMEGAMON® AI Insights trains themselves and makes predictions. Similarly, the frequency of the data received – hourly or daily – also shapes the context for the algorithm.

The following predefined contexts and corresponding configurations are available to optimize the machine learning engine.

Table 1. Context Names and Context Codes
Context name	Context code	Description
kmua-zos-hourly-sysplex	`kzhs`	Using the z/OS data collection (`km5`), you can detect anomalies on `MSU` consumption, major aggregation per Sysplex, and Service classes. This is an hourly process.
kmua-zos-daily-lpar	`kzdl`	Using the z/OS data collection (`km5`), you can detect anomalies on `MSU` consumption, major aggregation per LPAR, and Service classes. This is a daily process.
kmua-jvm-hourly-job	`kjhj`	Using the JVM data collection (`kjj`), you can detect anomalies on garbage collected, major aggregation per job. This is an hourly process.
kmua-network-hourly-lpar	`knhl`	Using Network data collection (`kn3`), you can detect anomalies on transmit segments, major aggregation per LPAR. This is an hourly process.
kmua-cics-hourly-region-cpu	kchrc	Using CICS data collection (kc5), you can detect anomalies on CPU time, major aggregation per Region. This is an hourly process.
kmua-cics-hourly-region-rt	kchrr	Using CICS data collection (kc5), you can detect anomalies on Response Time, major aggregation per Region. This is an hourly process.