Goal of a load testing activity is to identify and fix performance bottlenecks. Visual Studio Online integrates Load testing with Application Insights to give deep performance and diagnostics data of your application dramatically decreasing the discovery and resolution of any
issues discovered. This article will walk you through configuring Load Testing to take advantage of Application Insights.
*Please note that Application Insights is in Preview.
Here the steps to get started:
- Verify Cloud based load testing with Visual Studio Online is working correctly. Directions can be found here: http://aka.ms/loadtfs
- Make sure you have the latest Visual Studio Updates:
- Download Visual Studio 2013 Ultimate RTM
- Download Visual Studio 2013 Ultimate Update 2 RC
- Verify Application Insights is installed on the machine you want to load test. Please see: Get started with Application Insights.
Make sure your application(s) which you want to monitor are correctly configured with Application Insights and verify you can view the performance data flowing in the Visual Studio Online portal. Please note: Performance data in Application Insights can take more than 30 minutes to be visible in the Application Insights portal from the time you configure your application. If you don’t see your performance data in Application Insights please see: http://aka.ms/aimmafaq - Connect the Visual Studio IDE to your Visual Studio Online account from team explorer and open your Load Test.
- Add Application Monitoring to your load test, for monitoring during the load run.
a. Right click on Applications and choose the option of “Add Application for Visual Studio Online”
b. This will retrieve the list of Applications configured with your Application Insights.
c. Choose the Application(s) you want to monitor with the load run
6. Run the load test in Cloud.
7. Application Counters are correlated with the User load on the time axis. 
You can observe 6 counters for each Application. CPU, Memory, PerformanceEventsPerSec and ExceptionsPerSecond are the preselected ones. You can as well select NetworkIO and DiskIO and view them in the graph. These are correlated with User load on the time axis.
8. Go to Application Insights link and do a Root Cause analysis : To do the detailed root cause analysis of the issue, you can visit Application Insights with the help of a link in the Application Graph:
To learn more about how to do Root cause analysis of the performance issues with Application Insight, please visit: http://aka.ms/aigetstarted