Skip to main content

How to Measure Latency

Measuring latency for monitoring and benchmarking purposes is notoriously difficult. There are a lot of pitfalls with collecting, aggregating and analyzing latency data. In the talk, I make an effort to visit this topic from a top-down perspective and compile known complications and best-practice approaches on how to avoid them. This includes Measurement Overhead - Queuing effects - Coordinated omission - Histograms for Aggregation and Visualization - Percentile aggregation - Latency bands and burn-down charts - Latency comparison methods (QQ Plots, KS-Distance)