← Back to directory

Grafana MCP Server

Official

Query and manage Grafana dashboards, alerts, and observability data via MCP

GitHub source repository ↗

★ 3.3k Stars Category · Other Very popular

72FMRS · B

Reliability

11/20

Security and permissions

14/20

Maintenance

16/20

Documentation

18/20

Setup experience

13/20

mcp-grafana is officially maintained by Grafana Labs and offers broad coverage of the Grafana ecosystem — dashboards, alerting, many datasource types, incident management, OnCall, annotations, snapshots, and rendering — with fine-grained per-category tool toggles and a read-only mode. The main risks center on credential handling and scoping write access appropriately: configure least-privilege RBAC for the service account, avoid enabling write tools unless needed, and selectively enable only the datasource/tool categories actually in use to limit context usage and attack surface.

Read the FMRS scoring method →

mcp-grafana is the official Model Context Protocol server maintained by Grafana Labs, giving AI assistants access to a Grafana instance and its surrounding ecosystem. It covers dashboard search and editing, querying across many datasource types (Prometheus, Loki, InfluxDB, ClickHouse, CloudWatch, Elasticsearch/OpenSearch, Athena, Snowflake, Graphite, Quickwit, Pyroscope), alert rule and notification routing management, Grafana Incident management, Sift root-cause investigations, OnCall scheduling, annotations, snapshots, panel/dashboard image rendering, Agent Observability, Grafana Assistant, and provisioning-repository validation. Tools are grouped into categories that can be enabled or disabled via CLI flags, and the server can run in read-only mode with --disable-write.

Tools

list_teams

List all teams

list_users_by_org

List all users in an organization

list_all_roles

List all Grafana roles

get_role_details

Get details for a Grafana role

get_role_assignments

List assignments for a role

list_user_roles

List roles assigned to users

list_team_roles

List roles assigned to teams

get_resource_permissions

List permissions defined for a resource (dashboard, datasource, folder, etc.)

get_resource_description

Describe available permissions and assignment capabilities for a resource type

search_dashboards

Search for dashboards by title or other metadata

get_dashboard_by_uid

Retrieve full dashboard details using its unique identifier

update_dashboard

Update an existing dashboard or create a new one

get_dashboard_panel_queries

Get panel title, query, and datasource UID/type from every panel in a dashboard

run_panel_query

Execute one or more dashboard panel queries with custom time ranges and variable overrides (disabled by default)

get_dashboard_property

Extract specific parts of a dashboard using JSONPath expressions

get_dashboard_summary

Get a compact overview of a dashboard without the full JSON

list_datasources

List configured datasources

get_datasource

Get a datasource by UID or name

get_query_examples

Get example queries for a datasource type (disabled by default)

query_prometheus

Execute an instant or range PromQL query against a Prometheus datasource

list_prometheus_metric_metadata

List metric metadata

list_prometheus_metric_names

List available metric names

list_prometheus_label_names

List label names matching a selector

list_prometheus_label_values

List values for a specific label

query_prometheus_histogram

Calculate histogram percentile values (p50, p90, p95, p99)

list_incidents

List incidents in Grafana Incident

create_incident

Create an incident in Grafana Incident

add_activity_to_incident

Add an activity item to an incident

get_incident

Get a single incident by ID

query_loki_logs

Run log or metric queries using LogQL against Loki

list_loki_label_names

List all available label names in logs

list_loki_label_values

List values for a specific log label

query_loki_stats

Get statistics about log streams

query_loki_patterns

Query detected log patterns to identify common structures and anomalies

analyze_loki_labels

Audit a Loki label strategy (live or static) and optionally diagnose query performance

suggest_loki_alloy_label_config

Generate an Alloy loki.process snippet enforcing approved labels

query_influxdb

Query InfluxDB using InfluxQL (v1) or Flux (v2) (disabled by default)

list_clickhouse_tables

List tables in a ClickHouse database with row counts and sizes (disabled by default)

describe_clickhouse_table

Get column names, types, and metadata for a ClickHouse table (disabled by default)

query_clickhouse

Execute SQL queries with Grafana macro and variable substitution (disabled by default)

list_cloudwatch_namespaces

Discover available AWS CloudWatch namespaces (disabled by default)

list_cloudwatch_metrics

List metrics available in a namespace (disabled by default)

list_cloudwatch_dimensions

Get dimensions for filtering metric queries (disabled by default)

query_cloudwatch

Execute CloudWatch metric queries with time range support (disabled by default)

list_athena_catalogs

Discover available Athena data catalogs (disabled by default)

list_athena_databases

List databases in an Athena catalog (disabled by default)

list_athena_tables

List tables in an Athena database (disabled by default)

describe_athena_table

Get column names for an Athena table (disabled by default)

query_athena

Execute SQL queries against Athena with macro substitution and limit enforcement (disabled by default)

query_elasticsearch

Query Elasticsearch or OpenSearch using Lucene syntax or Query DSL (disabled by default)

query_quickwit

Query Quickwit using Lucene syntax or Elasticsearch-compatible Query DSL (disabled by default)

list_snowflake_tables

Discover Snowflake tables via INFORMATION_SCHEMA.TABLES, with optional database/schema filters (disabled by default)

describe_snowflake_table

Get column names, types, nullability, defaults, and comments for a Snowflake table (disabled by default)

query_snowflake

Execute SQL queries against Snowflake with macro and template variable substitution (disabled by default)

alerting_manage_rules

Manage alert rules (list, get, versions, create, update, delete)

alerting_manage_routing

Manage notification policies, contact points, and time intervals

list_oncall_schedules

List schedules from Grafana OnCall

get_oncall_shift

Get details for a specific OnCall shift

get_current_oncall_users

Get users currently on-call for a schedule

list_oncall_teams

List teams from Grafana OnCall

list_oncall_users

List users from Grafana OnCall

list_alert_groups

List and filter alert groups from Grafana OnCall

get_alert_group

Get a specific alert group from Grafana OnCall by its ID

get_sift_investigation

Retrieve an existing Sift investigation by its UUID

get_sift_analysis

Retrieve a specific analysis from a Sift investigation

list_sift_investigations

Retrieve a list of Sift investigations with an optional limit

find_error_pattern_logs

Detect elevated error patterns in Loki logs using Sift

find_slow_requests

Detect slow requests using Sift (Tempo)

list_pyroscope_label_names

List label names matching a selector

list_pyroscope_label_values

List label values matching a selector for a label name

list_pyroscope_profile_types

List available profile types

query_pyroscope

Query profiles, metrics, or both from Pyroscope

get_assertions

Get assertion summary for a given entity

agento11y_manage_conversations

List, search, and fetch LLM conversations from Grafana Agent Observability

agento11y_manage_generations

Fetch a single LLM generation and its evaluation scores

agento11y_manage_agents

Read the agent catalog: list agents, fetch a full agent version, view version history, and compare per-version score aggregates

agento11y_manage_evaluators

Manage evaluators, evaluator templates, and the judge catalog (list, get, upsert, fork, test, delete)

agento11y_manage_eval_rules

Manage eval rules that bind evaluators to production traffic, and inline guards (list, get, create, update, preview, delete)

agento11y_manage_eval_collections

Manage saved conversations and the collections that group them

ask_assistant

Send a natural-language prompt to Grafana Assistant and wait for the full text reply, with multi-turn support via contextId

generate_deeplink

Generate accurate deeplink URLs for Grafana dashboards, panels, and Explore views

get_annotations

Query annotations with filters (time range, dashboard UID, tags, match mode)

create_annotation

Create a new annotation, standard or Graphite format

update_annotation

Update specific fields of an existing annotation (partial update)

get_annotation_tags

List available annotation tags with optional filtering

list_snapshots

List dashboard snapshots with optional query and limit filters

get_snapshot

Retrieve snapshot metadata and dashboard payload by snapshot key

create_snapshot

Create a dashboard snapshot from a full dashboard payload

delete_snapshot

Delete a snapshot by snapshot key

get_panel_image

Render a stored dashboard/panel, or a provisioning preview from a repository branch, as a PNG image

list_provisioning_repositories

List provisioning repositories (e.g. git-sync sources) with source URL, branch, sync state, and health

validate_provisioning_file

Dry-run-apply a file from a provisioning repository and report admission validation errors

Setup

Install uv (which provides the uvx command). 2. In Grafana, create a service account and generate a token (or use the built-in Editor role for convenience, or configure minimal RBAC scopes for least-privilege access). 3. Add an mcpServers entry to your MCP client config (e.g. Claude Desktop, Cursor) with command 'uvx', args ['mcp-grafana'], and env vars GRAFANA_URL and GRAFANA_SERVICE_ACCOUNT_TOKEN (Grafana Cloud users replace GRAFANA_URL with their instance URL, e.g. https://<instance>.grafana.net). 4. Alternatively run the Docker image docker.io/grafana/mcp-grafana over stdio, or connect directly to the hosted remote endpoint https://mcp.grafana.com/mcp (streamable-http, with an optional X-Grafana-URL header to target a Grafana Cloud instance). 5. Use --enabled-tools or --disable-<category> flags to control which tool categories are exposed, and --disable-write to run in read-only mode.

claude_desktop_config.json

{"mcpServers":{"grafana":{"command":"uvx","args":["mcp-grafana"],"env":{"GRAFANA_URL":"http://localhost:3000","GRAFANA_SERVICE_ACCOUNT_TOKEN":"<your service account token>"}}}}

Generate and validate client config

Fit and risk

Best for

Teams running a Grafana instance (self-hosted or Grafana Cloud) who want an AI assistant to help with observability workflows
DevOps/SRE engineers who want to automate querying alerts, logs, and metrics as part of incident response
Scenarios where monitoring data access needs to be exposed to an AI assistant safely in read-only mode

Not for

Teams with no Grafana ecosystem in place (no Prometheus/Loki/etc. datasources configured)
Fully offline or network-isolated environments (the remote endpoint and most datasource queries require network access)
Situations where write access must be tightly scoped but the team is unwilling to configure fine-grained RBAC (the built-in Editor role, while convenient, is broad)

Required permissions

GRAFANA_URL (required): the target Grafana instance address
GRAFANA_SERVICE_ACCOUNT_TOKEN or GRAFANA_USERNAME/GRAFANA_PASSWORD: authentication credentials (the service account token is sensitive)
Grafana RBAC permissions matching the enabled tool categories, e.g. dashboards:read/write, datasources:query, alert.rules:read/write, annotations:read/write, or the built-in Editor role as a simpler but broader alternative
Grafana Incident and Sift tools use basic roles (Viewer for reads, Editor for writes) rather than fine-grained RBAC
Optional GRAFANA_ORG_ID or X-Grafana-Org-Id header for multi-organization support
Optional GRAFANA_EXTRA_HEADERS and GRAFANA_FORWARD_HEADERS to add or forward HTTP headers (e.g. Cookie) to the Grafana API

Risks and side effects

Service account tokens, passwords, and other credentials are supplied via environment variables or headers and must be safeguarded from leakage
Write tools (update_dashboard, alerting_manage_rules, create_incident, create_annotation, create_snapshot, etc.) can directly modify or delete dashboards, alert rules, and incident data — misuse can damage production configuration
Using the built-in Editor role for convenience grants broad access beyond least-privilege
Fetching full dashboard JSON (get_dashboard_by_uid, update_dashboard) can consume significant context window space and may expose sensitive queries or variables embedded in the dashboard
GRAFANA_FORWARD_HEADERS forwarding of headers like Cookie, if misconfigured, risks leaking session information
The hosted remote endpoint https://mcp.grafana.com/mcp is a third-party managed service — using it to reach a Grafana Cloud instance routes requests through that service
SSE/streamable-http transport misconfiguration (e.g. --allowed-hosts or --allowed-origins set to * while exposed on an untrusted network) carries DNS-rebinding-style risk

Troubleshooting

Verify GRAFANA_URL is correctly formatted and reachable; Grafana Cloud users should use https://<instance>.grafana.net
If a tool call returns a permission error, check that the service account token's RBAC permissions and scopes cover the operation, or temporarily test with the Editor role to isolate a permissions issue
For datasource-related failures, confirm the Grafana instance is version 9.0 or later
If an expected tool is missing, check whether its category is disabled by default (e.g. admin, clickhouse, cloudwatch, snowflake) and add it via --enabled-tools
A 403 on SSE/streamable-http requests usually means the Host or Origin header isn't covered by --allowed-hosts / --allowed-origins
Use the --debug flag to enable detailed HTTP request/response logging when diagnosing connectivity issues
If token rotation isn't taking effect in Kubernetes, confirm GRAFANA_SERVICE_ACCOUNT_TOKEN_FILE (not an inline token) is used and that the mounted Secret has actually updated
If get_panel_image fails, confirm the Grafana Image Renderer service is installed and configured on the target instance

Use cases

AI-assisted incident triage: query Prometheus/Loki metrics and logs, run Sift root-cause analysis, and create a Grafana Incident

Natural-language dashboard discovery and lightweight editing without manually hunting through the Grafana UI

Unified querying and exploration across many datasource types (Prometheus, Loki, ClickHouse, CloudWatch, Athena, Snowflake, and more)

Automatically generating deeplinks to specific dashboards, panels, or Explore views for reports and tickets

Supported clients

Claude DesktopFull support

CursorFull support

Related MCP servers

Compare with other MCP servers