How AI Performance Testing Works

TABLE OF CONTENTS

• AI is Coming to the Rescue!
  • AI Application Discovery
  • AI Test Design and Execution
  • Performance Insights

AI is Coming to the Rescue!

If you've read through our thoughts on performance testing, you've learned that the test engineers who specialize in the PSR approach are very opinionated. At Testaify, we are equally opinionated. If your application cannot meet specific requirements, we will tell you. Sorry, but you have an ugly baby; we might even show a photo of Kramer (kidding, we do not want to pay royalties). As we've covered, performance testing covers several quality attributes: responsiveness, scalability, stability, and reliability.

An advantage of Testaify is that we will discover your system in a secure way, before your customers find any issues. We will know the different transactions and the navigation of your system. We will learn your domain. Using that information, Testaify can design your performance tests, execute them, and tell you whether your baby is ugly.

While our first product will focus on functional testing, the second will focus on performance. Still, the underlying platform is the same for all our products. As such, it is important to discuss the Testaify discovery process. It is an essential component of our platform.

AI Application Discovery

Testaify’s discovery uses AI to navigate your application and build a model of it. It does this every time you execute a test run. We know how your product evolves as you add features and fix defects. That data will allow us to identify potential response time issues immediately.

Our performance testing team used to discuss the idea of a single-user test with engineers at Ultimate Software. If you are developing a new page or modifying a page, you can easily do a manual test and time the response time for a single user. Essentially, we are trying to help the engineers shift-left performance testing. Only a few engineers engage in that effort.

Testaify Performance for Web will do this automatically as part of every run. We will have an early warning of where a performance problem might occur. While discovering the application, we can identify specific pages with increased response time. Users can review those performance warnings before the test run starts.

AI Test Design and Execution

After we build the model of the application, Testaify Performance for Web will design a set of tests to determine the system's scalability. The model will inform our system, which will develop a set of performance test scripts to create a comprehensive suite.

Inspired by our experience at Ultimate Software, we will execute simultaneous user tests against critical pages. Our AI system will choose the targets for that type of testing. It will also design a concurrent user test scenario against the overall system.

Our AI workers can generate different load levels to determine the application's scalability threshold. The amazing serverless architecture built by our excellent engineering team will allow us to find that threshold quickly. We can horizontally scale as needed to complete the test.

One component we will not be able to control is the test environment, as it is provided by our customers when they register the specific application instance.

Performance Insights

Testaify Performance for Web will provide their users with a clear picture of the response time of every page at a single user level. It will tell the customers what is the scalability threshold of the application. It will provide insights into the overall performance of the system. It will show the long-term trend for every page in the system and provide an early warning if specific pages are getting slower or becoming error-prone during performance testing.

As we expand the capabilities of Testaify Performance for Web, we will integrate our data with customers’ APM systems. That will open up the opportunity to provide more information about the other quality attributes of the application. The combination of the test data and production data will allow us to provide forecasts regarding the system's scalability, stability, and reliability. It will also enable Testaify Performance for Web to refine the test suite for every test run.

And, yes, we will tell you if the baby is breathtaking like Elaine.

Stay tuned as we move on to discuss Usability in future blog posts.