VisualPlan displays visual execution plans that are used to execute queries. For more information about how to troubleshoot query performance issues, see Slow query table.
You can use VisualPlan by stage. The lower part of the flowchart is the most upstream stage. The most upstream stage reads and filters data from source tables. Arrows indicate the directions in which data flows. Data processing logic in a stage is represented by a combination of operators. Arrows indicate the directions in which data flows between operators.
- The total values of major metrics for all tasks in the stage. The following metrics
- Output: the total amount of output data and the total number of output rows in the stage.
- Stage: the ID of the stage, which is the same as that on the details page.
- State: the final status of the stage.
- Operator Cost: the total CPU time consumed by all threads of all tasks in the stage to execute a query.
- Peak Memory: the peak memory used by the stage to execute a query.
- Input: the total amount of input data and the total number of input rows in the stage.
- Estimated information
- Est. I/O: the estimated number of input and output rows of an operator. If a Join operator is used, only the number of output rows is estimated. In the execution plan optimization phase, a specific method is used to estimate the amount of data for an operator whose number of output rows is uncertain. This way, a better execution plan can be selected. If the estimated and actual values are significantly different from one another, inappropriate execution plans may be selected. In this case, a large amount of server resources may be consumed in queries or slow queries may occur.
- For more information about operators, see Operators.