SELLEST OSAST
📘 Free e-book: The 7 success factors of software testing. 25 years of project experience in one 33-page workbook, now also in English 👉 Get it for free
"Performance testing is not the same as load testing." - Leandro Melendez
In this episode, I talk with Leandro Melendez about how performance testing changed in the last 20 years. Live at HUSTEF, we swap stories from bare metal and heavy browser scripts to APIs, cloud, and Kubernetes. Leandro draws a clear line between performance and load testing. Do not run Black Friday tests every sprint. Watch production, use canaries, and learn from real users. He pushes observability first. Build dashboards, instrument early, and think about cost.
Leandro is helping everyone to ramp up on their observability, QA and performance practices. He has over 20 years of experience in IT and over 15 in the performance testing practice where he served multiple S&P500 customers all over the USA, Mexico, Canada, Brazil, India, Austria, etc. Author of the performance testing blog Señor Performo where he curates a diverse set of learning material for performance testers and engineers together with a couple of YouTube channels in Spanish and English. As well he hosts the PerfBytes español Podcast, and is a co-host on the main PerfBytes show since 2018. He is an international public speaker participating in multiple conferences, events and webinars, with keynotes, workshops and multiple talks on his belt. And last, author of The Hitchhikers Guide To Load Testing Projects, a fun walkthrough that will guide you through the phases or levels of an IT load testing project.
Highlights:
- Running a load test every sprint makes no sense: continuous releases with small changes call for observability and monitoring, not repeated capacity tests designed for peak events.
- Performance testing and load testing are not the same thing; treating them as identical is a habit left over from waterfall projects and bare-metal servers that no longer applies to modern cloud systems.
- Elastic cloud infrastructure removes the hard capacity ceiling but creates a cost problem: a system that auto-scales without being tuned can spend far more than necessary while still performing poorly.
- AI systems carry a performance cost that is easy to underestimate now because token prices are low, but poorly designed AI integrations with large context windows will become expensive as pricing normalizes.
- Observability, meaning agents, telemetry, and a human-readable dashboard, is the starting point for performance work on any project, not the choice of a load-testing tool.
