← Back to the chapter · Previous detail
Technical detail · Time-Series Anomalies: Baselines and Residuals
Baseline drift and seasonal mismatch
Seasonal baselines become stale when equipment, firmware, occupancy, weather, or operating policy changes. Adaptive methods can reduce false alerts and also absorb a slow fault into the baseline. Separate drift monitoring from incident scoring and require baseline review after known site changes, maintenance, firmware updates, schedule changes, or sustained drift.
When slow degradation stops alerting after the baseline catches up, suspect an update rate that is too fast or unreviewed training windows. Freeze baselines during incidents and monitor drift separately from alert scoring.
When normal weekend or holiday behavior causes many alerts, suspect missing temporal context or the wrong seasonal period. Use context fields, calendar rules, and residual review by operating mode. Baseline review triggers retesting rather than assuming a changed pattern is healthy.