Datadog Anomaly Detection Agent
Monitors Datadog metric streams using the Datadog API v2 and applies ML-based anomaly detection to alert on infrastructure drift. Integrates with PagerDuty and Slack webhooks for multi-channel incident routing.
Monitors Datadog metric streams using the Datadog API v2 and applies ML-based anomaly detection to alert on infrastructure drift. Integrates with PagerDuty and Slack webhooks for multi-channel incident routing.
npx skills add agentskillexchange/skills --skill datadog-anomaly-detection-agent-2
The Datadog Anomaly Detection Agent provides intelligent monitoring by connecting directly to the Datadog Metrics API v2. It continuously evaluates time-series data for CPU, memory, disk I/O, and custom metrics, applying statistical models (DBSCAN, Z-score, seasonal decomposition) to identify anomalous patterns before they escalate into incidents.
Key capabilities include automatic threshold calibration based on historical baselines, multi-metric correlation to reduce false positives, and configurable severity levels. When anomalies are detected, the agent routes alerts through PagerDuty Events API v2 for on-call escalation and Slack Block Kit messages for team visibility.
The agent supports tag-based filtering to scope monitoring to specific services, environments, or teams. It can also generate Datadog dashboard JSON definitions for visualizing detected anomalies alongside normal operating ranges. Configuration is YAML-driven with support for per-service override profiles.