Doing or Waiting?

This article emphasizes the importance of understanding whether application performance bottlenecks stem from “doing” (CPU usage) or “waiting” (external dependencies or resource contention). By using observability tools like APM and distributed tracing, developers can pinpoint what an application is waiting for – be it direct requests to other services or indirect waits due to CPU limitations like garbage collection – and address the root cause. Ultimately, optimizing for both efficient processing and minimized waiting leads to improved performance and cost savings, as demonstrated by the example of optimizing SQL queries instead of simply adding more hardware.