So you’ve decided that it’s time to load test your website or app? Awesome. There are a probably a handful of questions that are swirling around your head as you prepare to do some performance testing.
- What metrics do I need to get out of the test?
- Do I want to test locally or from the cloud?
- Should I test with concurrent or parallel requests?
- How many virtual users do I need to simulate?
That last question in particular has come to my attention numerous times in previous weeks. While the rest of these questions are necessary to consider, the number of virtual users you simulate can vary drastically per organization. Deciding on your virtual user quota will likely be a pivotal factor in whether or not you get the most possible ROI out of your load tests.
I recently sat down with Michael Punsky, SmartBear’s internal load testing expert, to get his insight into how someone should figure out precisely how many virtual users they need to simulate. Punsky, who has supervised more than 900 load tests throughout his career, was able to answer this question (and touch upon the decision of whether or not to test from the cloud) in less than two minutes. Hopefully his words of wisdom will help make this decision a little bit easier for you as you embark on your next performance testing saga:
About Michael: Michael J. Punsky is the Manager of Professional Services for SmartBear. Michael has over 30 years of experience in technology in both the private sector and theUnited States Air Force. He specializes in the delivery of services, and mainly focuses his writing on website performance and load testing. You can find a list of articles written by Michael on the Software Quality Matters blog.
- Synthetic and Real-User Monitoring: Better Together
- GeekSpeak with Scott Barber: Alan Page
- What Is Load Testing?