grafana

Overview

The grafana MCP server is a Model Context Protocol (MCP) server that enables AI assistants and agents to interact directly with Grafana dashboards, data sources, and observability resources through a structured, AI-friendly interface. It allows AI-driven workflows to explore metrics, logs, alerts, and dashboards without switching tools or manually navigating the Grafana UI.

This server is ideal for observability, incident response, performance analysis, and operational workflows where real-time telemetry and monitoring context are essential.

Transport

sse

Tools

• add_activity_to_incident
• create_alert_rule
• create_annotation
• create_folder
• create_graphite_annotation
• create_incident
• delete_alert_rule
• fetch_pyroscope_profile
• find_error_pattern_logs
• find_slow_requests
• generate_deeplink
• get_alert_group
• get_alert_rule_by_uid
• get_annotation_tags
• get_annotations
• get_assertions
• get_current_oncall_users
• get_dashboard_by_uid
• get_dashboard_panel_queries
• get_dashboard_property
• get_dashboard_summary
• get_datasource_by_name
• get_datasource_by_uid
• get_incident
• get_oncall_shift
• get_sift_analysis
• get_sift_investigation
• list_alert_groups
• list_alert_rules
• list_contact_points
• list_datasources
• list_incidents
• list_loki_label_names
• list_loki_label_values
• list_oncall_schedules
• list_oncall_teams
• list_oncall_users
• list_prometheus_label_names
• list_prometheus_label_values
• list_prometheus_metric_metadata
• list_prometheus_metric_names
• list_pyroscope_label_names
• list_pyroscope_label_values
• list_pyroscope_profile_types
• list_sift_investigations
• list_teams
• list_users_by_org
• patch_annotation
• query_loki_logs
• query_loki_stats
• query_prometheus
• search_dashboards
• search_folders
• update_alert_rule
• update_annotation
• update_dashboard

Key Capabilities

• Dashboard exploration — Browse and inspect Grafana dashboards and panels programmatically.
• Metrics and telemetry access — Query time-series data from connected data sources such as Prometheus, Loki, or other backends supported by Grafana.
• Alert visibility — Inspect alert rules, states, and related metadata for operational awareness.
• Incident analysis — Enable AI assistants to correlate dashboards, metrics, and alerts during troubleshooting.
• Observability-driven workflows — Embed monitoring context directly into AI-driven investigations and summaries.

How It Works

Check out our Grafana MCP server guide.

The grafana MCP server runs as an MCP service that connects to a Grafana instance using configured credentials and permissions. AI clients communicate with the server over the MCP protocol to request observability context as part of broader reasoning workflows.

The server mediates access to Grafana’s APIs, handling authentication, query execution, and response normalization before returning structured results that AI assistants can reason over. This abstraction ensures that all access respects Grafana’s organization, folder, and permission models while shielding clients from API-specific details.

By exposing Grafana observability data through MCP, the server enables AI workflows such as summarizing system health, investigating anomalies, and supporting on-call engineers — all driven through natural language and AI reasoning within a single environment.