August 2026
On reporting intervals
Why a fixed sampling interval flatters slow processes and penalises bursty ones.
Short write-ups. Most run under two pages.
Why a fixed sampling interval flatters slow processes and penalises bursty ones.
Comparison baselines that quietly move, and a cheap check that catches most cases.
What we do when the sample is too small for the recommended method.
A taxonomy of double-counting we keep running into, and how each one shows up.