Baseline
The historical reference point used to determine what normal looks like for a signal.
title: "Baseline" description: "The historical reference point used to determine what normal looks like for a signal."
A baseline is the expected range or pattern for a metric based on prior observations. Early warning systems rely on baselines to separate meaningful deviations from ordinary variation.
Why Baselines Matter
Without a baseline, every alert threshold becomes arbitrary. A value that is risky in one environment may be routine in another.
Typical Inputs
- Historical observations
- Time windows
- Reference models such as mean and standard deviation
- Freshness and staleness rules
In Practice
Baselines allow teams to ask a better question than "did this number go up?" The more useful question is "did this number move outside the range we would reasonably expect?"