no. of requests made per second with N users

LegacyForumLegacyForum Posts: 1,669 ✭✭
edited December 2016 in Load Test
The attached graph generated (detailed report) after my load test. I have selected a scenario, which contains three unit test clients, and numbers of users is the controlled parameter here. The N users concurrently invoked the scenario for 5 minutes of duration. "Test Completion Rate (1/sec)" is at y-axis and x-axis the "Test Completion Time (sec)". My questions are..

1) In the first raised point is 5 Hits. Can we consider that there are 5 Hits per second at beginning of test and 8 Hits per second at the end of test?
2) What is ?point average? scale option in the graph?
3) How the graph will help us to find out how many hits are made per second while N users (say 10 users) concurrently called the test clients?
4) If say 5 Hits Per Second, that mean 5 transactions (request + server process + response back to client) completed in one second or just 5 requests are reached at server end?
5) If I have RunCount (say 100) and Test Duration or Test Completion Time (say 100 seconds), then can I assume that 100/100, i.e 1 transaction per second has happened?


  • LegacyForumLegacyForum Posts: 1,669 ✭✭
    1) The green graph line shows the tests completion rate. The 8 at the end of the load test indicates that there was a sudden increase in the test completion rate at the end of the test.

    2) The point average scale option is useful when there are a lot of points and you want to smooth the graph. The point average scale option determines how many points to use to calculate the average. For instance, if it was set to 5, it would take the first 5 points, find the average, plot that as one point, take the next 5, find average, etc.

    3) If you want to look at the hits per second, then you should enable the checkbox for Tests Start Rate (1/sec)

    4) 5 hits per second means that SOAtest sent 5 requests per second.

    5) 1 transaction per second would be the average. However, the test completion time for each individual test could be very different from the average. For instance, the average of (10, 10, 10) is 10, however, the average of (1, 1, 28) is also 10.
Sign In or Register to comment.