Notes / Grafana Cloud / Reference

What is Grafana MCP

Grafana's official open-source MCP server (mcp-grafana) — gives AI agents tool access to query metrics/logs/traces, manage dashboards and alert rules, and work with Incident and Sift; also available as a hosted OAuth 2.1 remote server with 50+ tools.

Updated July 9, 2026 · §202607081949-4 ·

Grafana MCP (grafana/mcp-grafana) is Grafana Labs’ official Model Context Protocol server. It gives any MCP-aware agent (Claude, an ADK agent, an IDE assistant) tool access to a live Grafana instance — the same category of thing as What is MCP Toolbox, but for observability data instead of databases.


What it exposes

CapabilityConcretely
Metrics & logsQuery Prometheus/Mimir and Loki through Grafana’s datasource layer
DashboardsSearch, read, and manage dashboards
Alert rulesCreate, read, and manage alerting
Incident & SiftWork with Grafana Incident and Sift investigations directly
DeeplinksGenerate links back into Grafana for a human to open the same view
Agent (Claude / ADK / IDE)


   mcp-grafana server  ──▶  Grafana instance
                                 ├── Mimir  (metrics)
                                 ├── Loki   (logs)
                                 ├── Tempo  (traces, via a dedicated tracing MCP server)
                                 ├── Alerting
                                 └── Incident / Sift

Two deployment shapes

Self-hosted (open source) — run it yourself via uvx mcp-grafana, pointed at your instance:

GRAFANA_URL=https://your-stack.grafana.net \
GRAFANA_SERVICE_ACCOUNT_TOKEN=glsa_xxxxxxxx \
uvx mcp-grafana

Hosted (public preview, 2026) — Grafana now runs a remote MCP server at mcp.grafana.com/mcp, authenticated with OAuth 2.1 over Streamable HTTP, exposing 50+ tools: everything in the open-source server plus Grafana Assistant capabilities. No self-hosting required, at the cost of routing your observability queries through Grafana’s hosted endpoint rather than your own process.

Token type: don’t reach for the wrong one

The MCP server authenticates to the Grafana HTTP API with a service-account token (glsa_…) — this is a different credential and a different code path from the glc_-prefixed access-policy tokens used for Alloy’s remote-write into Mimir/Loki/Tempo. Mixing those up is a known 401 trap on the write path; for the MCP server specifically, glsa_ is correct because you’re calling the Grafana API, not writing metrics/logs directly to the data plane.

Cardinality/blast-radius note

Because execute_sql-style generic tools are risky in What is MCP Toolbox, the equivalent consideration here is scope of the service account token. A token with dashboard-edit and alert-rule-write permissions handed to an agent means the agent can modify production alerting — scope the service account to the minimum the agent actually needs (read-only for query/triage agents; write scope only for agents explicitly meant to remediate).

Why it’s on the backlog: it’s the direct enabler for extending the h-aiops SRE-agent line beyond read-only dashboards into agent-driven triage — pair with What is Grafana Skills to standardize how the agent uses these tools, not just that it can.

Local graph

Full graph →

Linked from 10 notes

What is MCP Toolbox

Google's open-source MCP server for databases (formerly Gen AI Toolbox for Databases) — production-ready (v1.0.0) prebuilt tools that connect agents and IDEs directly to AlloyDB, Spanner, Cloud SQL, BigQuery, and more, with OAuth2 zero-trust gating.

What is gcx

Grafana's official CLI for managing Grafana Cloud/Enterprise/OSS resources, optimized for agentic usage — dashboards, alerts, SLOs, metrics/logs/traces/profiles queries, and every major Cloud product, plus a bundled Agent Skills set for Claude Code and other .agents-compatible harnesses.

What is Grafana Skills

Grafana's take on reusable agent skills — captures a team's troubleshooting expertise as standardized, shareable procedures that can trigger MCP actions, plus gcx: a Grafana Cloud CLI shipping 22 bundled skills for Claude Code and other .agents-compatible harnesses.

What is Envoy

CNCF-graduated L7 proxy built at Lyft — the de facto data plane for service mesh (Istio, Linkerd's predecessor lineage) — now extending into AI traffic via Envoy AI Gateway, which reached v1.0 with a native MCP Gateway in 2026.

Agentic AI: Projects & Engineering Mastery

A book-shaped table of contents for Agentic AI: Projects & Engineering Mastery: hands-on practitioner builds, Principal/Staff-level technical leadership, and the lookup appendices and vendor/framework reference notes for the whole series. Book 6 of the AI Systems Engineering series.

What is Harness Engineering

The discipline of designing everything that wraps a raw LLM into a reliable agent — tool contracts, system prompt architecture, context/memory management, permission gates, and feedback loops. Distinct from prompt engineering and model training.

What is HolmesGPT

Robusta.dev's open-source SRE agent (CNCF Sandbox) for investigating production incidents across Kubernetes, VMs, cloud services, and databases — an agentic tool-calling loop over 70+ toolsets, not a chatbot or RAG system, with a proactive 'operator mode' that monitors and opens fix PRs without a human trigger.

What is Vertex AI

Google Cloud's managed ML/AI platform — as of 2026 rebranded and consolidated into the Gemini Enterprise Agent Platform, bundling 200+ foundation models, Agent Builder, and a managed agent runtime (formerly 'Agent Engine').

Grafana Cloud

A book-shaped table of contents for Grafana Cloud: platform foundations through telemetry collection, Mimir/Loki/Tempo/Pyroscope, visualization, application observability, reliability tooling, developer experience, governance, and enterprise reference architectures — cross-linking existing notes instead of duplicating them.

What is Telegraf

InfluxData's plugin-driven metrics/events/logs collection agent — 300+ input/output plugins, written in Go, single static binary — the collector layer in the InfluxDB (TICK-stack-descendant) ecosystem, comparable in role to Grafana Alloy.