Struct InsightsEngine 

pub struct InsightsEngine { /* private fields */ }

The ML traffic insights engine (DW-092): in-process, hand-rolled EWMA forecasting and seasonal-baseline anomaly detection. Held inside crate::analytics::EmbeddedAnalytics behind an Arc and fed by the live-sketch window rotation.

Implementations

impl InsightsEngine

pub fn new(config: &AnalyticsInsights) -> Arc<Self>

Construct from the insights config. baseline_windows is clamped to at least 1 (the ring buffer must hold at least one window).

pub fn forecast_enabled(&self) -> bool

Whether forecasting is enabled.

pub fn anomaly_enabled(&self) -> bool

Whether anomaly detection is enabled.

pub fn baseline_windows(&self) -> u64

The configured baseline window cap.

pub fn observe(&self, window: BaselineWindow)

Observe one completed sketch window: add it to the recent ring buffer (evicting the oldest when full) and update the minute-of-day seasonal entry with an exponential moving average.

pub fn forecast(&self, now_ms: i64) -> ForecastResult

Forecast the next window’s shape: the seasonal average for the next minute-of-day, trend-adjusted by an EWMA over the recent ring. Confidence is the fraction of the seasonal entry’s data volume relative to a full day of observations (capped at 1.0).

pub fn detect_anomaly(&self, current: &BaselineWindow) -> AnomalyResult

Detect whether the current window is anomalous relative to the seasonal baseline for its minute-of-day. A metric is flagged when the current value exceeds the baseline by a factor of 2.0 (the ANOMALY_FACTOR constant). The anomaly score is the maximum current/baseline ratio across the checked metrics, normalized to [0, 1].