chevron-thin-right chevron-thin-left brand cancel-circle search youtube-icon google-plus-icon linkedin-icon facebook-icon twitter-icon toolbox download check linkedin phone twitter-old google-plus facebook profile-male chat calendar profile-male
0 votes
We are trying to set up automated load tests as part of the build.
The idea is that we can monitor the performance of business components and have metrics that give us an indication of code changes affecting the performance.
We assembled load tests by selecting from our existing integration tests.
Unfortunately this did not work as it seem that tests using Typemock failed more or less randomly when executed in a Load test setting.

Arnulf
asked by krok (3.9k points)

3 Answers

0 votes
Arnulf,
This is a very interesting situation. Could you tell us how you run the load test and collect metrics. Are you using a 3rd party application to run the load tests? In any case, if you post the failing tests, it might help.
answered by eli (5.7k points)
0 votes
Don't forget that Typemock Isolator uses the Profiler API which can affect the performance of your application. (This is true for any profiler - try running, say, the CLR Profiler on your app, then running the app without it - you'll notice the performance difference.)

The reason I bring it up is that the results of running your app with a profiler going can be potentially misleading.
answered by tillig (6.7k points)
0 votes
We use the Load test feature in Visual studio.
It seems like any test using mocking will fail.
Using the integration test for load testing will not give us exact metrics.
The mocking framework and fixture setup and teardowns will influence the results. But by comparing several executions we can eliminate these and we will be able to detect any potential decrease or increase in performance. At least that is the idea.
Arnulf
answered by krok (3.9k points)
...