Anomaly detection EDI
On this page
Anomaly detection for EDI flows, AI & data integration applied to EDI.
Definition
EDI anomaly detection applies ML models (Isolation Forest, Autoencoder, LSTM, Prophet) on flow metrics: volumes per partner/hour, amount distributions, mapping latencies, error rates. Detects drifts that would pass under static thresholds — typically a 30% drop in INVOICs from a retailer during holidays.
Origin
ML discipline formalised in the 2000s, applied to observability by Datadog Watchdog, Dynatrace Davis from 2018.
Example in context
Metric daily_invoice_count_partner_X monitored by Datadog Watchdog: auto-alert if deviation > 3σ.