</>MCP Agents Market
MCP Server

mcp-grafana

by grafana3.4kGoUpdated 2026-09-07

MCP server for Grafana

Claude DesktopCursorVS Code

The Grafana MCP server is an official Model Context Protocol server that connects AI assistants to Grafana instances for observability and monitoring workflows. It enables natural language interaction with dashboards, metrics, logs, traces, alerts, incidents, and other Grafana resources through tools for querying Prometheus, Loki, CloudWatch, SQL datasources, and managing alerts, incidents, and on-call schedules. The server supports both local Grafana and Grafana Cloud deployments, with configurable tool sets, RBAC permission scoping, and multiple transport modes (stdio, SSE, HTTP). It's particularly useful for SREs and developers who want AI-assisted troubleshooting, dashboard creation, and incident response through conversational interfaces like Claude Desktop.

Key Features

Query multiple datasources: Prometheus (PromQL), Loki (LogQL), CloudWatch, Elasticsearch, InfluxDB, SQL databases (ClickHouse, Snowflake, Athena, MySQL, PostgreSQL, MSSQL), Graphite, and Pyroscope
Dashboard management: search, read, create, update, get summaries, extract properties with JSONPath, and execute panel queries with custom time ranges
Alerting and incident management: create/update alert rules, manage notification policies, contact points, Grafana Incident workflows, and Grafana OnCall schedules
Advanced observability: Sift investigations for error patterns and slow requests, Agent Observability for LLM conversation analysis, and Grafana Assistant integration
Granular RBAC permission support with per-datasource, per-dashboard, and per-folder scopes for least-privilege access
Multi-organization support with dynamic per-call org selection, service account token file rotation, and custom HTTP header forwarding
Query guardrails for Loki with configurable byte limits and time-range caps to prevent expensive queries
Comprehensive observability: Prometheus metrics, OpenTelemetry distributed tracing, OTLP log export, and slow-request logging

Use Cases

  • 01AI-assisted incident triage: ask Claude to find error patterns in Loki logs, correlate with slow Tempo traces, and create a Grafana Incident ticket
  • 02Natural language dashboard exploration: query 'show me CPU usage for prod over the last hour' without writing PromQL
  • 03Automated alert rule creation: describe alert conditions in plain English and let the AI generate and deploy the alert rule configuration
  • 04On-call workflow automation: check current on-call users, acknowledge alert groups, and update incident status through conversational commands
  • 05Cross-datasource investigation: correlate metrics from Prometheus, logs from Loki, and traces from Tempo in a single AI-guided troubleshooting session
  • 06Dashboard generation: describe desired visualizations and have AI create or update dashboards with appropriate panels and queries

Related MCP Servers

View more

markitdown

Python tool for converting files and office documents to Markdown.

MCP Server188k

firecrawl

The context API to search, scrape, and interact with the web at scale. 🔥

MCP Server187.9k

prompts.chat

f.k.a. Awesome ChatGPT Prompts. Share, discover, and collect prompts from the community. Free and open source — self-host for your organization with complete privacy.

MCP Server171.9k

langflow

Langflow is a powerful tool for building and deploying AI-powered agents and workflows.

MCP Server155.5k

mcp-grafana — FAQ

What is the Grafana MCP server?+

It's an official Model Context Protocol server that connects AI assistants like Claude Desktop to Grafana instances, enabling natural language queries for metrics, logs, dashboards, alerts, and incident management. It supports both local Grafana and Grafana Cloud deployments.

How do I install and configure it?+

Install via uvx (recommended), Docker, or download the binary. Add configuration to your MCP client (e.g., Claude Desktop's config JSON) specifying the command (uvx mcp-grafana or docker run), Grafana URL, and service account token. The server requires Grafana 9.0+ for full functionality.

What clients does it work with?+

It works with any MCP-compatible client including Claude Desktop, Cursor, and VS Code (using stdio, SSE, or streamable-http transports). The server supports multiple transport modes to accommodate different client architectures.

What credentials or API keys are required?+

You need a Grafana service account token with appropriate RBAC permissions for the tools you want to use, or username/password credentials. For Grafana Cloud, use your instance URL (e.g., https://myinstance.grafana.net) and a service account token. The README provides detailed RBAC scope examples for each tool.

Is it free and open source?+

Yes, it's open source under the Apache 2.0 license and free to use with both Grafana OSS and Grafana Cloud. Some advanced features like Agent Observability and Grafana Assistant require specific Grafana Cloud plugins.

Can I restrict which tools are available?+

Yes, you can enable/disable entire tool categories using flags like --disable-write, --disable-query, --enabled-tools, or disable specific categories like --disable-oncall, --disable-sql. This helps manage context window usage and enforce security policies.

How do I install mcp-grafana?+

Open the source repository on GitHub and follow its README. mcp-grafana is a mcp server — MCP Agents Market links you directly to the official repo.

Is mcp-grafana free?+

mcp-grafana is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches

Analytics & Monitoring MCP ServersDeveloper Tools MCP ServersCloud & DevOps MCP ServersBest SkillsBest PluginsBest AgentsTop rated listingsClaude Code toolsBrowse all categories