Load Impact’s basic functionality is geared towards simulating user behavior on your site in the most realistic way possible.
In a real-life situation, if you were to have concurrent users on your website at the same time, you probably won’t have all of them on the same page doing precisely the same things. This is why you can run a test with up to 10 different user scenarios. This allows you to create a variety of realistic user scenarios(i.e. test scripts) and have them all play out at once.
There are two ways to generate a realistic load script using Load Impact:
1.Google Chrome Extension (recommended)
- Follow these steps to use the Chrome Extension
When you start a recording, a new browser window with your target URL opens, and whatever actions you do there will be recorded.
- Follow these steps to use the Proxy recorder
The HTTP requests will then be translated to a script when the recording is over, which you can further edit if necessary.
What happens when I actually start the test?
So when you start a test, what actually happens is that we will start by running your script with one user, and slowly ramp up to the number of VUs you configured the test for. When the script runs to completion, it will re-run itself until the time runs out.
Now if you look at your script, you will see that there is something called “client.sleep”. That is the time that the user spends browsing each page, and serves as a simulation of what real users do.
Since we add clients to your site in a ramp up situation, what you will get is that some users would be on Page 3 of your script while others are just starting on Page 1. This will mean that each page will never be subject to a load of X number of concurrent users loading resources at the same time.
However, it is important for you to know what your testing goals and objectives are prior to the scripting of the test. If your test objectives are to simulate a marketing campaign where you expect most of your traffic to flood to one page. In this case, you might want to dedicate one user scenario which simulates the repetitive loading of one or two pages.
What's the difference between an "auto-generated" load script and the one I record with the Chrome Extension or Proxy Recorder?
Testing a single URL though the "auto-generated" function will not capture the complexity and mix of behaviors your users are likely to exhibit in real-life. The script generated through the auto-generate function will only load objects from the target URL or page (i.e. images, files, HTML), and will not capture the pattern of behavior.