Why you don't need to worry about scaling your Java webapp

A conference talk walkthrough on how to approach load testing for Java web applications without over-engineering for scale. The speaker demonstrates a practical methodology using multiple EC2 loaders, a probe machine, and a server running embedded Jetty. Key concepts covered include: separating loaders from the probe to get reliable latency measurements, collecting OS-level metrics (CPU, memory, network I/O) for all participants, using flame graphs and GC logs for detailed profiling, and interpreting HDR histograms. The talk emphasizes that most applications will never need to handle the extreme request volumes developers fear, using Stack Overflow's ~300 req/s as a real-world reference point. It also stresses the importance of testing with realistic user journeys and real infrastructure rather than single-endpoint limit testing.