Datadog MCP Server

Executes Datadog metric queries using the native Datadog Query Language (DQL) through the MCP protocol, translating natural language requests into structured metric API calls. Supports aggregation functions, time-range specifications, and multi-metric comparisons by parsing user intent and constructing properly-formatted Datadog API requests that return time-series data points with timestamps and values.

Lists and retrieves detailed configuration of Datadog monitors (alert rules) including thresholds, notification channels, and current alert status. Implements pagination to handle large monitor inventories and filters monitors by type (metric, log, APM, synthetic) and status (triggered, ok, no data) by calling the Datadog monitors API endpoint and parsing the response into structured alert rule objects.

Searches logs stored in Datadog using the Datadog Log Query Language, supporting field-based filtering, boolean operators, and faceted aggregations. Translates natural language search intents into structured log queries, handles pagination of large result sets, and returns log entries with parsed fields, timestamps, and source metadata. Implements facet extraction to enable drill-down analysis on specific log attributes.

Retrieves distributed traces and individual spans from Datadog APM, supporting filtering by service, operation, trace ID, and span tags. Constructs trace queries using Datadog's trace query syntax and returns hierarchical span data including timing, error status, and custom tags. Enables correlation between traces and other observability signals (metrics, logs) through shared trace IDs and service names.

Creates, updates, and retrieves Datadog dashboards through the MCP interface, supporting widget configuration (graphs, tables, heatmaps), layout management, and dashboard templating. Translates high-level dashboard specifications into Datadog dashboard JSON schema, handles widget positioning and sizing, and manages dashboard permissions and sharing settings through API calls.

Retrieves events from Datadog's event stream, including monitor alerts, deployments, and custom events, filtered by time range, source, and tags. Reconstructs incident timelines by correlating events with metrics and logs, enabling chronological analysis of system state changes. Supports event aggregation and deduplication to identify related incidents.

Queries Datadog's tag infrastructure to discover hosts, services, and metrics by tag filters, enabling dynamic resource inventory and dependency mapping. Returns tagged resource lists with metadata (host status, service dependencies, metric availability) and supports hierarchical tag queries (e.g., 'env:prod AND service:payment-api'). Enables agents to dynamically identify relevant resources without hardcoded resource lists.

Executes Datadog synthetic tests (API, browser, multi-step) and retrieves test results including response times, error details, and assertion failures. Supports on-demand test execution and polling for test completion, returning detailed failure information for debugging. Enables agents to validate service availability and functionality programmatically.

Integrates with Datadog's incident management and service catalog APIs to retrieve incident details, service ownership, and on-call schedules. Enables agents to identify incident responders, understand service dependencies, and correlate incidents with service metadata. Supports incident status updates and escalation through the MCP interface.

Retrieves Datadog billing and cost data, including per-service costs, usage metrics (ingested logs, indexed spans, custom metrics), and cost trends. Enables cost attribution by service or team and supports cost forecasting based on historical usage patterns. Helps identify cost optimization opportunities by analyzing usage anomalies.