Performance issues come in all flavors... - CPU usage (complexity, cache misses...) - Memory usage (leaked resources like memory, connections and handles) -

Performance issues come in all flavors... - CPU usage (complexity, cache misses...) - Memory usage (leaked resources like memory, connections and handles) - Excessive I/O (networking, files...) - Concurrency (contention, starvation, deadlocks, livelocks...) There's no one size fits all profiler. I find load testing in combination with profiling very useful to get an idea of how a system performs. Usually I keep injecting load until the system fails then find what failed. There are other approaches, like automated performance testing (writing tests with assertions on performance). e.g: https://scalameter.github.io https://scalameter.github.io