How Anomalies Are Detected #

The Anomaly_Detector class runs as part of the nightly metric calculation (2:00 AM WP-Cron). It checks each monitored metric for deviations from baseline.

Detection Algorithm #

For each metric:

Step 1: Calculate baseline
The baseline is the metric’s rolling average over the configured lookback window (default: last 30 days).

Step 2: Calculate standard deviation
The standard deviation measures how much the metric normally varies day-to-day.

Step 3: Check current value
Compare today’s metric value to the baseline.

Step 4: Calculate deviation
Deviation % = (Current Value – Baseline) / Baseline x 100

Step 5: Apply threshold
If the absolute deviation exceeds the configured threshold (default: 20%), an anomaly is flagged.

Severity Thresholds #

These thresholds are configurable in Analytics > Settings > Anomaly Detection.

Metrics Monitored #

By default, anomaly detection monitors:
– MRR (daily)
– Churn rate (daily)
– Download volume (daily)
– License activation count (daily)
– Revenue (daily)

You can add or remove metrics from monitoring in Analytics > Settings > Anomaly Detection.