User guide

The Avg statistic shows the average execution time for the step across data slices, and the percentage
of the total query runtime that represents. The Max statistic shows the longest execution time for the
step on any of the data slices, and the skew. The skew is the difference between the average and
maximum execution times for the step.You might want to investigate a step if the maximum execution
time is consistently more than twice the average execution time over multiple runs of the query, and if
the step also takes a significant amount of time (for example, being one of the top three steps in
execution time in a large query).
Note
When possible, you should run a query twice to see what its execution details will typically be.
Compilation adds overhead to the first run of the query that is not present in subsequent runs.
To investigate high skew for a step, check the query plan for distribution steps to see what type of
distribution is being performed in the query, then review your data distribution strategy to see if should
be modified. For more information about Amazon Redshift data distribution, go to Choosing a Data
Distribution Style in the Amazon Redshift Database Developer Guide.
You can click any bar in the chart to compare the data estimated from the explain plan with the actual
performance of the query, as shown following.
If the query optimizer posted alerts for the query in the STL_ALERT_EVENT_LOG system table, then
the plan nodes associated with the alerts are flagged with an alert icon.
API Version 2012-12-01
190
Amazon Redshift Management Guide
Viewing Query Performance Data