Intelligence

Anomaly Detection

Predictive alerts with statistical baselines, anomaly scoring, and automated response actions.

Start Free Read the Docs

Capabilities

Complete anomaly detection toolkit

Statistical Baselines

Automatically compute mean, standard deviation, P95, and P99 baselines using sliding window analysis.

Deviation Scoring

Real-time anomaly scores based on deviation from baselines. Configurable sensitivity thresholds.

Six Metric Types

Monitor request rate, error rate, latency, token usage, score changes, and data volume.

Severity Levels

Automatic severity classification from info to emergency based on deviation magnitude.

Auto-Actions

Configurable responses: alert, throttle, suspend, or kill switch based on anomaly severity.

Baseline Comparison

Side-by-side baseline vs current value comparison with trend analysis and forecasting.

Developer Integration

Detect anomalies programmatically

anomaly.ts

import { DRD } from '@drd/sdk';

const drd = new DRD({ token: 'drd_live_sk_...' });

// Get anomaly summary
const summary = await drd.anomaly.getSummary();
console.log(summary.active);              // 3
console.log(summary.averageAnomalyScore); // 0.78

// List active anomalies
const anomalies = await drd.anomaly.list({ severity: 'critical' });

// Acknowledge and resolve
await drd.anomaly.acknowledge(anomalies[0].id);
await drd.anomaly.resolve(anomalies[0].id);

Detect before it becomes an incident

Statistical baselines, real-time scoring, and automated response actions.

Get started free View documentation