Recently, when doing performance test analysis of LR results, the problem of different response time values of each transaction in summary and average transaction response time was also encountered. Make a note here.
If you set the thinking time in the transaction, you need to be aware of the filter thinking time.
Set whether the method is included in the View->summary filter, whether there is a filter that contains the think time (the last item in LR11).
Summary: The default is to sample based on the elapsed time of the entire scene. If you need to modify it, set the execution time of the scene in the View->summary filter.
Average transaction response time: LR automatically sets the sampling time based on factors such as scene run time. You can view the sample time interval for the current view by selecting the set granularity item in the view's right-click menu.
therefore, Summary and the average transaction response time graph is the largest, the average figure, the minimum figure, the reason is different because the sampling is different. If the sampling time is different in two views, there is no contrast.
Summary is the average of the time of the whole scene, the maximum minimum value, is also taken from the entire scene.
In the average transaction response time graph, the values are taken by frequency.
Similarly, there is a problem with sampling time in the controller.
Configuration of sampling time in controler see
Reference:
1, Lr_ problem _ Average Response time interpretation, summary and analysis inconsistent
2.
Interpretation of the average transaction response time for LR analysis and the different time values in summary