Configuring a load test with multiple user scenarios

We recently had a great question come in from one of our customers that we thought we would share.

Question: Planning to run a test with 10.000 concurrent users spanning 4 or 5 user scenarios.  How do I configure a test to run with, say 35% of the load running user scenario 1, 35% running user scenario 2, 10% running user scenario 3 etc?

And, when running multiple scenarios, where each scenario consists of 2 or more pages, how can we see the performance (load time) of each page in each scenario?

Answer: Assigning a certain amount of the simulated users to each user scenario is something you do in the “Test configuration” section.

Just scroll down the page to the section called “User scenarios”, then click the “Add scenario” button to add a new user scenario to the test. When you have all the scenarios you want added, you can fiddle with the percentages to get the exact load allocation for each scenario that you want.

 

The load time of each page in a user scenario can be collected if you use the –

http.page_start() and http.page_end() functions

– inside the user scenario script. Read more about that here and here.

Example: page metrics

-- Log page metric
 http.page_start("My page")
 responses = http.request_batch({
    { "GET", "http://loadimpact.com/" },
    { "GET", "http://loadimpact.com/style1.css" },
    { "GET", "http://loadimpact.com/image1.jpg" },
    { "GET", "http://loadimpact.com/image2.jpg" }
 })
http.page_end("My page")

Using the above script as a user scenario would result in a plot-able page load time metric for a page called “My page”. The name of the page can be changed to whatever you want.