Questions about data-driven load tests

2 posts, 0 answers
  1. Michael
    Michael avatar
    1 posts
    Member since:
    Sep 2015

    Posted 02 Sep 2015 Link to this post


    I'm a developer on a web-based application that has the following requirements for a load test:
    1) Log in as one of 100 preset users
    2) Navigate to a dynamic URL based on values in a table in a database
    2) Process at that URL

    Among other things. However, I'm struggling with retrieving the data from the database and also creating that URL. The URL is of the format /Site/Controller/Function/ID, the same as almost any MVC application, but I see no place where I can choose a datasource for the original URL (e.g., Further uses find the dynamic target correctly, but I need to change that original ID in the first request after login. The ID is normally selected via a Telerik grid, but I also can't figure out how to get the link from the grid either.

     I'm also struggling with the lack of multiple data sources. At this point, I feel as though I have to merge my two data sets (users and data values) and query from that. However, I have concerns about multiple VU's overwriting each other, which isn't a true test, and leads me to my next question:

    Assuming that I use a database-driven test, when is that query executed? Every time a new VR is created? When the test first runs and never again? Every time a variable is retrieved?

    Thank you for your assistance with this.

  2. Cody
    Cody avatar
    3354 posts

    Posted 03 Sep 2015 Link to this post

    Hello Michael,

    Provided your login credentials are sent to the webserver via some web form or query string we can data drive that. Unfortunately the URL itself cannot be data driven. Test Studio load tests can only work with static pre-recorded URL's. We can only data drive the parameters of the query string and web form data.

    Test Studio load tests cannot draw data from two separate sources. What happens is during test initialization all the data is read and an in memory .NET DataTable object is created out of the data. The test is then run using that data. The closest we can come to using multiple sources is a T-SQL query that uses a JOIN clause.

    Assuming that I use a database-driven test, when is that query executed?

    What query exactly are you referring to here? As stated above, all of the data is read during test initialization then the test uses that data. During test execution each VU consumes the next available row of data. When the entire table has been consumed, it is recycled, i.e. the next VU will get row 1, and the next VU will get row 2 and so on. This keeps going until the time runs out on the test.

    I have concerns about multiple VU's overwriting each other,

    Can you clarify what "overwriting each other" means to you in the context of a load test?

    The New Release of Telerik Test Studio Is Here! Download, install,
    and send us your feedback!
Back to Top