Method for Detecting Skew - Teradata Workload Management - Analytics Database

Teradata Vantageā„¢ - Workload Management User Guide - 17.20

Deployment
VantageCloud
VantageCore
Edition
VMware
Enterprise
IntelliFlex
Product
Analytics Database
Teradata Workload Management
Release Number
17.20
Published
June 2022
ft:locale
en-US
ft:lastEdition
2023-01-27
dita:mapPath
aji1628112479055.ditamap
dita:ditavalPath
qkf1628213546010.ditaval
dita:id
B035-1197
Product Category
Teradata Vantage

In general, TASM detects skew successfully using just the skew percentage metrics. However, some situations can lead to a false detection of skew. In a multi-step request, one step could be very skewed but also short and not indicative of an overall problem with the request. It is important with skew detection that temporary skew within a request not be interpreted as persistent skew and that temporary skew be ignored.

Tip: Do not set the qualification time too low on skew exceptions. The skew condition should persist long enough to qualify as a problem with the request.