16.10 - States and Events Overview - Teradata Workload Management

Teradata Database Workload Management User Guide

prodname
Teradata Workload Management
vrm_release
16.10
created_date
June 2017
category
Administration
User Guide
featnum
B035-1197-161K
The best practice under normal operating conditions is to use only one ruleset. Instead of creating multiple rulesets for different situations, you can create states to define how TASM allocates resources during different times of the day, different days of the month, or when system performance is degraded. Moving from one state to another, due to planned or unplanned circumstances, results in an automatic change in the workload management settings. Only one state in one ruleset can be in effect at a time. State definitions consist of these factors:
  • Health conditions: These define system health, for example, a degraded health condition. You can define an unplanned event that triggers a health condition, such as a down node.
  • Planned environments: These are time frames or different processing windows that the system supports, for example, a Late-Night planned environment for batch jobs.
Administrators define events that cause a health condition or a planned environment to change. A change to either or both causes a state change, depending on how the state matrix is defined. Events include the event definition and the actions that Teradata Database takes when the event occurs. The event definition can be based on the occurrence of a single event or a combination of events. When an event occurs, Teradata Database can take one of these actions:
  • Change the health condition or planned environment
  • Send an alert
  • Run a program
  • Post a notification to the system queue table (DBC.SystemQTbl)

Once you create a state, you can apply filters, throttles, and workloads based on that state.