Continuous Performance Testing: A Tale of Two Worlds
Once upon a time, in a land far away, application performance was a differentiator in the market. But in today's competitive landscape, having a highly responsive and scalable application is a basic expectation. Performance no longer lives on the fringe of the quality world, and it is becoming more common for organizations to include performance testing as part of their CI/CD pipelines. However, engineers are faced with a common dilemma: Do they implement small and simple performance tests that may not be very realistic but are fast and reliable in the build pipeline, or do they write comprehensive, end-to-end performance tests that are representative of the real world but result in slower builds? Join Kaushal Dalvi as he discusses his journey to get closer to the answer to this and other performance testing questions. Discover the challenges and solutions around continuous performance testing, and learn how to strike the right balance so that you can design better performance tests for your pipelines. Don't miss the chance to hear how Kaushal and his team have been navigating performance testing in the worlds of monoliths versus microservices, UI versus API tests, on-premise solutions versus public cloud, mobile versus desktop, and proprietary versus open source tools.
Since he got his first computer at age thirteen, Kaushal Dalvi has been interested in systems and software performance. He spent days researching performance characteristics of different motherboards, CPUs, GPUs, RAM, and disks to configure and overclock them in order to squeeze the maximum frames per second out of the games he played. Kaushal built and maintained websites for local businesses, where he started learning about performance and reliability. He developed a taste for it, and now ten years later he continues working in the same field as the director of quality and performance engineering at Ultimate Software, leading teams that deal with workload modeling, test scripting and execution, bottleneck analysis, debugging, and much more.