A commonly used load testing model is to keep transaction rates constant regardless of how loaded the server system is already.
In
Silk Performer you can define workloads based on transaction or session rates, which
Silk Performer tries to meet even when server response times increase due to the increasing load. Consequently, when server response time
slows,
Silk Performer automatically reduces think times or pacing wait times to keep session times constant.
Silk Performer offers two methods to keep transaction or session rates at a defined level.
- Think time adaptation modifies the think times in your script so that a certain session time is reached. This option is only applicable if the
virtual user script contains at least one
ThinkTime() function call.
- Pacing wait time is the period of time that a virtual user idles after a transaction or session has been completed, so that a certain transaction
or session rate is reached.
To modify pacing options, click
Run Test on the workflow bar and click the
Pacing tab.
To open the
Configure Pacing dialog, click
... in the
Pacing column.
Several options are available to configure pacing:
- Think time adaptation: Click this option to determine a
Goal Session Time. The think times will be automatically adjusted to match the specified goal session time. You can select between
Static and
Dynamic. Think time adaptation will only work as expected if the virtual user script contains at least one
ThinkTime() function call.
- Wait time insertion: Specify either the
Pacing wait time or the
Goal Session Time. If you specify the pacing wait time, it will be added to the execution time of a transaction. If you add up these two times,
you will get the goal session time. If you specify the goal session time, the pacing wait time will be automatically determined.
You can also specify to randomize the entered values.
To configure pacing wait time directly in a BDL script, use the BDL functions
SetPacingTime() and
GetPacingTime().
Silk Performer offers pacing for the following workload models:
- Increasing
- Steady State
- All Day
- Dynamic