Get the event-volume anomaly snapshot
CDP
Get the event-volume anomaly snapshot
Compare the most recent complete hour of cdp_events ingest volume against a learned baseline (mean and standard deviation of the preceding hourly buckets) and classify it as spike, drop, normal, or insufficient_data. Use it to catch a broken SDK deploy, an expired ingest key, a mis-rotated HMAC secret, or a client-side retry storm early — these show up as a sudden drop or spike in ingest volume long before they surface as tracking-plan violations. New or trickle-volume tenants return insufficient_data and never alert. Owner / admin / developer / marketer.