Load Testing, Reframed: Test User Journeys, Not Just APls

​We load test. We monitor. We alert. So why do performance issues still reach production?

​A server can boast 99.9% uptime and still deliver a terrible user experience. Why?

​Because users don't care if an API returned a 200-they care whether they actually completed their purchase.

​The problem is that we're often measuring the health of individual services, not the success of the user journey.

​This talk flips the traditional load-testing mindset. Instead of asking, "Can the service handle the traffic?", we ask, "Can users complete their journey?"

​We'll stress-test the flows that matter most, the experiences that make or break your product.

​Behind every "simple" user action in microservices lies a web of interconnected services. Different teams, different dependencies, different failure modes-one weak link can send shockwaves across the entire journey.

​We'll revisit one of our favorite battle scars: a performance issue that every load test missed, but our users didn't.

​Finally, we'll compare how today's popular load-testing tools-k6, Gatling, Locust, and JMeter-approach journey-based testing, and which ones are best suited for testing what users actually experience