In this article, you are going to learn about the importance of Correlation in Performance Testing. Correlation is a critical process in performance test scripting. It doesn’t matter which tool you are using for performance testing. You can ignore correlation only if you are testing direct pages e.g. if you are testing only the home page load, contact page load, pricing page load etc., then you cannot perform correlation.
Correlation is important when you are dealing with session IDs, date and time in milliseconds, dynamic values in the URLs, responses, headers etc.
What is Correlation in Performance Testing?
Correlation is a process of capturing and storing the dynamic response from the server and passing it in the subsequent requests.
How to identify correlation in performance testing?
Identifying correlation can be done in multiple ways. Below are the usual methods to identify correlation in performance testing.
Experience plays a vital role in correlation. Experienced performance testers or software developers under test can easily identify the correlation by simply parsing the script. They will be able to quickly identify directly from the recorded script e.g. identifying strings such as session, customer ID, policy ID, view state, event state, date and time in milliseconds etc.
Recording it twice and comparing
Another effective but time-consuming method is recording the action twice with the same set of data and comparing it side by side.
- Record Script A and Script B
- Open Script A and Script B in LoadRunner
- Right click on Script A’s Action in Solution Explorer and select File to Compare
- Right click on Script B’s Action in Solution Explorer and select Compare to <script path> as shown below.
LoadRunner utility called WDiff compare the action of Script A and B and provides you the conflicts in a highlighted yellow as shown below.
From the Logs
You can identify the correlation steps from the logs generated in LoadRunner. E.g. you can get the dynamic values from the replay log and compare it with Recording log.
From the snapshot pane
By comparing the page view of each step of recording and replay as shown below. This is one of the easiest ways to identify the dynamic values. But it might be time-consuming if there are more resources for each request.
Please share your experience of correlation and let me know your way of doing correlation in the comments section.