Datadog AI Beginner
Datadog offers some of the most accessible AI-powered monitoring features. This lesson covers Watchdog, anomaly monitors, forecast monitors, and network-specific AI features for network operations.
Watchdog
Watchdog is Datadog's automatic anomaly detection engine. It monitors all your metrics and logs without any configuration, surfacing anomalies as "Watchdog alerts" in a dedicated feed.
- Automatically detects metric anomalies across infrastructure, APM, and logs
- No manual baseline configuration needed — it learns patterns automatically
- Provides root cause analysis with correlated events
- Particularly useful for network metrics like interface utilization spikes and error rate changes
Anomaly Monitor Types
| Algorithm | Description | Best For |
|---|---|---|
| Agile | Fast-adapting, adjusts to recent changes quickly | Metrics with recent pattern shifts |
| Robust | Ignores recent anomalies when calculating baseline | Stable metrics with occasional spikes |
| Adaptive | Adjusts to seasonal patterns dynamically | Metrics with strong daily/weekly patterns |
Network Performance Monitoring (NPM)
Datadog NPM provides flow-level visibility into network traffic with AI-enhanced analytics:
- Network map — Automatic topology visualization of service-to-service communication
- DNS monitoring — Track DNS resolution times and failure rates with anomaly detection
- Retransmits and latency — Per-connection TCP metrics with baseline comparison
Network Device Monitoring (NDM)
NDM collects SNMP metrics from routers, switches, and firewalls with built-in anomaly detection and out-of-the-box dashboards for popular vendors (Cisco, Juniper, Arista, Palo Alto).
Forecast Monitors
Create monitors that alert when a metric is predicted to cross a threshold in the future. For example, alert 48 hours before a disk fills up or a link reaches capacity.
Next Step
Learn how to configure Splunk ITSI for service-level network monitoring.
Next: Splunk ITSI →