Need help with StresStimulus? Start here.
Not getting the results I expect
|
D Rain Posts: 1
7/4/2013
|
First off, I have no knowledge in this area at all, so I may be using StressStimulus incorrectly. My requirement is basic, does the new version of the software perform as well as the old.
Testing a new version of vendors ASP .net web app via IE. When I visually run the new version in IE against the old version, the new version takes twice as long to display all the contents.
To gauge the difference, I decide to use StressStimulus but it always gives me the result that the new version performs better.
Example: Run record and open one URL I want to test and then stop once it finishes loading. Fiddler lists about 900 requests. I run a steady load, VU of 3, test duration of 5 mins, wait for response and warm up at 1 sec. Run the test. I close FIddler down, do teh same again with old version. The new version processes 550 iteration and the old version processes 60 in 5 mins.
I make a change to Test duration of number of iteration per Vo to 5 and run and the new version runs in les taht half the time
Why is it the case when i compare manually in IE I can perform 10 operations (button click etc) in old version before I can perform 5 operations in the new yet teh stresStimulus has it teh other way around?
Any help will be appreciated.
|
|
|
0
link
|
|
Vadim @StresStimulus Administrator Posts: 583
7/4/2013
|
Hello,
Without seeing your actual tests, I cannot answer your question on why the tests complete faster in the new version. There can be many reasons for that. Here are a few examples.
Test in the new version require some parameterization to run correctly and because it is missing, the server returns an error message instead of executing all actions performed in the old version. Another example is that new version has a faster server and slower client, meaning that it completes rendering the page on a client using slow JavaScript. StresStimulus, as all load testing tools, only measures the server and network time, but not the browser DOM time as the latter is not impacted by a high user load.
If you upload both tests to our secure support repository, I can examine and let you know why you're getting such result. To do so, record one URL in the old version, save the test as Test-Old, run the test with Test duration = 5 iteration per VU. After test is complete, in the StresStimulus main menu -> Support Options -> Upload Support Case, check all checkboxes and click Upload. Then record one URL in the new version, save the test as Test-New, run and upload the test in the same way.
Cheers,
-Vadim
|
|
|
0
link
|
Copyright © 2024 Stimulus Technology