The V2 release switches towards supporting and using Fiddler as the default tool for recording the requests that are made, Fiddler is a mature tool that supports recording requests made over https

To Create a new project with a new test please start Fiddler and capture the http/https requests. once the requests are recorded you can remove the unnecessary requests and save all the requests using the File->Save->All Sessions menu action

Create a new project and save it.

Right Click on the Test Definition node and select New Test Definition->Web Test Wizzard

Using the Import Fiddler Data import the data that was saved earlier. Modify the name of the test and click on Generate Test Definition

The Test Definition source code editor window will open, there go to the compile tab and click compile code. Once the code compiles succesfully click on the "Save C# Source Code" button.

After you have defined a test definition you can create a test run definition that will use the previously defined test definition. The test run definition allows you to set how the previously define test will be ran.

To define a new test run right click on the Test Run Definition nodse and select new Test Run Definition. Than set up all the properties as in the picture below.

Well not much :).

The html content is there just to help you prepare the test script/program. So you will have to check each page and see that all your user actions that you have recorded in the prowser are happening propperly here too. There are a lot of scenarios where due to the way the web application was done load testing it will require a little tinkering with the C# code (ex the developers have put a functionality for protecting against robots)

After you have a woking test script/program you will have to decide on the Pause between requests and start load testing the web application for a specific amount of time lets lay 10 minutes with 1 thread. After those ten minutes you check the results and if everything is ok you should increase the number of threads to 5 and after that to 10 threads, 15 ... and so on until the web application that you are testing stops responding propperly. You can also check the CPU and memory consumption on the web server, if the CPU stays constant at 100%(or a lower value for real servers) during the test you should proably stop because you probably reached your web application+Web server limit.

When you find the number of threads for which the web application doesn't work propperly it means that the number of users that web application will support = Number of Threads