16.10 - Averaging Qualification - Teradata Workload Management

Teradata Database Workload Management User Guide

Teradata Workload Management
June 2017
User Guide

The metrics associated with some events fluctuate wildly. To detect this type of event, use the Averaging qualification method. This method requires that the moving average of the metric exceed the specified threshold.

To understand the need for Averaging, consider the following scenario. ResUsage data is collected at 10-minute intervals. However, event qualifying criteria are collected at every event interval, which can be 5, 10, 30, or 60 seconds. The following figure shows a graph of CPU use measured on an actual production system with 30-second collection intervals. As you can see, CPU use varies greatly. Compare that to the moving averages in the trend lines at 5 minutes, 10 minutes, and 30 minutes. The smoothing effect of moving averages helps a DBA get a clearer picture of actual CPU use. If a DBA wants to detect if CPU use falls under 75% for at least 10 minutes, Simple qualification does not activate the event. CPU use fluctuates too much for that. However, Averaging qualification does activate the event.

Use Averaging to qualify these system event types:
  • CPU Utilization
  • CPU Skew
  • I/O Usage
Use Averaging to qualify these workload event types:
  • CPU Utilization
  • Arrivals
  • SLG Response Time (for workloads created with SLGs)
  • SLG Throughput (for workloads created with SLGs)