The author draws a Hamlet-inspired analogy to criticize a common statistical error in tech monitoring: averaging percentiles across servers to produce misleading latency graphs. The piece highlights how this practice is widespread in SRE/observability workflows despite producing bogus, unstable numbers, and laments the general lack of statistical literacy among engineers.
Background
The article references a well-known Hacker News discussion about engineers averaging percentiles in monitoring dashboards, a practice that produces mathematically invalid results. This is a recurring topic in the SRE and observability community.
- Source
- Lobsters
- Published
- Jul 31, 2026 at 10:35 PM
- Score
- 5.0 / 10