To assign performance levels to measures,
Silk Performer needs to know the measures that will be available after a load test. Therefore, make sure to run a baseline test and set
a baseline before you follow the steps described below. If no baseline is set, there are no measures available to assign performance
levels to.
-
In the Workbench, click
.
-
The
Define Performance Levels dialog displays, containing a table. The table rows represent the performance levels. You can add as many performance levels
as you require.
-
The columns represent the performance indicators. Per column, select one of the following performance indicators:
- Avg: average value
- P50: median
- P90: 90th percentile
- P95: 95th percentile
- P99: 99th percentile
-
Name your performance levels and enter values for each level.
-
When your performance levels are defined, click
Assign Levels.
-
The
Assign Performance Levels dialog shows all measures that were collected during the baseline test. You can assign the defined performance levels manually
to each measure. Or you can click
Automatically set levels based on baseline. You can also filter the list to quickly find a specific measure, and you can click
Reset all levels to none. When you are done, click
OK.
Complete modeling your load test and execute it. When the execution is completed, open an overview report and look out for
the performance levels. They are color-coded: Performance levels that are met display in green, performance levels that are
missed display in red. This allows you to instantly see whether a load test is to be considered as passed or failed. To tweak
your settings after the load test execution, follow the reactive approach.
Note: A performance level definition is valid for a particular workload. If you have multiple workloads, you can define performance
levels for each workload.