Measuring End-to-End Latency

You will monitor the full request cycle to detect if errors are actually slow successes caused by system bottlenecks.