Daily builds need daily testing. In continuous engineering, you develop many releases. You need to test all these releases. Continuous engineering requires continuous testing.
Performance is an important aspect of your application. So, you want to test the performance. Low-intensity performance testing could be useful for you. In low-intensity performance testing, you sample only the response times.
If you test daily, you want to see which build shows a different performance. Performance Tester provides a report that provides visual insight.
Performance testing, what do you mean?
If you say performance testing, you can mean different types.
For an encyclopedic description of performance testing, jump to the Wikipedia article of Software performance testing.
Kind of Smoke Testing
However, you just might not have the people and time to do all these tests. So, what problems do you expect and how much time do you want to invest?
A fair tradeoff could be sampling only the response times before every release. You start a single-user test before every new release and watch the response times.
The response times say something about the first impression how an application performs under load. A different response time is likely to give different results in a real performance test. In fact, all performance tests consist of a set of response times.
Maybe you can compare sampling the response time with smoke testing.
Take an informed decisionThrough daily testing of your build, you can immediately pinpoint the build where performance issues occur or start to occur. With the HTTP trending report of Rational Performance Tester, you can discover deviations from the trend in a glance. See the picture.
Now, you can take the opportunity to fix and to defend your reputation.
Do you want to go live with your release or do you need to wait for a while? For you to make an informed decision.
Run the Test Daily
For a low number of users, you can use the Rational Performance Tester Starter Edition. For testing the response times of a daily build, you can put just a single user in a test. Also, a schedule of five or less users could be an option.
In short, you can think of the following work.
In the graphs, you see that on one day the response is much slower. Also, the page health shows a drop. In that sense, you are dealing with an outlier. Now, it’s time to figure out what’s wrong with the build. In this example, the development team fixed the issue. The next build shows normal values again.
Share Your Experiences
In continuous engineering, you want to see the performance for every build. If a build shows strange response times, you know that something is wrong in the build. Be fruitful and fix.
If you’d like to share your experiences with low-intensity performance testing, leave a comment here. Follow me on Twitter @gunangwaney.
Gunang Waney works as a product specialist of the Rational Testing tools. He is passionate about communicating technical knowledge and making things work for customers. Gunang graduated from Utrecht University, The Netherlands with an MSc degree and works in Amsterdam, The Netherlands. When he can escape from his computer, he spends time with his family.