Introducing anomaly detection in Datadog

Anomaly detection has been added to Datadog to provide deeper context for dynamic metrics like application throughput, web requests, and user logins. The feature analyzes a metric's historical behavior to distinguish between normal and abnormal trends. It accounts for seasonality and can separate the trend component from the seasonal component of a timeseries. Anomaly detection is available in Datadog and complements outlier detection, which identifies unexpected differences in behavior among multiple entities reporting the same metric.