anomaly-detection-rule-builder

Anomaly detection in PostgreSQL uses SQL templates for static thresholds, statistical deviations, and trend detection across business metrics. It includes drop-in rules, an anomaly_log, and an investigation playbook without requiring external dependencies.

Statistical deviations can be monitored across revenue, traffic, conversions, churn, and operational metrics. Detection rules operate over daily to monthly horizons to identify significant anomalies in these specific business areas.

Yes, the anomaly detection rules are PostgreSQL-compatible and explicitly designed to be Supabase and PostgreSQL friendly. They function in standard SQL environments using only standard libraries, requiring no external dependencies.

False positive alerts during known business events are handled through event-aware suppression. This feature prevents automated alerting when expected metric fluctuations occur, ensuring investigations focus only on genuine anomalies.

The anomaly investigation playbook provides structured procedures for responding to detected metric deviations. It accompanies the SQL templates and anomaly_log to guide proactive alerting and systematic post-detection analysis.