When most people think of performance testing, they think about the hard parts – the very hard parts. They think about the expensive and complicated tools that are required to simulate the activity of thousands of end-users all at the same time, while collecting tens or hundreds of thousands of measurements. In fact I’ve frequently told those looking to break into performance testing to start by becoming a “mid-level everything”. Once when I mentioned to a friend that I was considering writing a how-to style book on performance testing, he quipped “You’re going to fill 150 pages with the phrase ‘Hire a consultant’ in a bunch of different languages?”
While there is at least some truth behind each of those statements, many performance issues can be detected and diagnosed with exactly the tools and knowledge you have at your disposal right now using information obtained from quick, easy and cheap performance tests. In fact, much of the performance related information that stakeholders need to make good decisions and development teams need to dramatically improve system performance is easily obtainable by the performance-testing layman. During this keynote, Scott Barber will introduce you to several techniques that the performance testing layperson can use to speed up and simplify the collection of valuable performance-related information; many of which can be accomplished in 30 seconds or less.
Conducting high quality performance testing is hard. It is hard to obtain and configure environments correctly. It is hard to install and configure monitoring and diagnostics tools. It is hard to collect valid and useful performance goals. It is hard to model the projected activity of hundreds or thousands of users accurately. The one thing that should not be hard is automating the performance test once you’ve overcome all of those challenges, but it often is. There are certain things that one must consider carefully before and during the automation of performance tests to avoid having it become another item in the list of things that are hard about performance testing.
In this class, Scott Barber shares 10 Must-Know Tips for Performance Test Automation that he has gleaned from his years of experience as a performance test consultant and coach. These tips are not based on theory alone, they have been applied by Scott, his clients, his peers, and his students and have been widely shown to increase automation value, reduce the effort of automating, or both, in a variety of environments with multiple automation tools to achieve diverse testing goals. While these tips are valid for any test automation effort, when it comes to automating performance tests they are essential if your goal is to keep automation off the list of things that are hard about conducting performance testing.
Main Message:
- Bad performance test automation is worse than useless, it yields misleading results.
- Good performance test automation demands significant knowledge of the SUT, as well as the business goals and expectations of the SUT.
- Bad performance test automation is easy to create, difficult to detect, and more difficult to correct.
- 10 tips to help you avoid bad performance test automation.
