Grafana MCP server provides access to Grafana instances and their surrounding ecosystem through the Model Context Protocol. Developed by Grafana Labs' Machine Learning team, it enables searching dashboards, fetching datasource information, querying Prometheus metrics, and managing incidents in Grafana Incident. The server supports both stdio and Server-Sent Events (SSE) transport modes, can be run locally or deployed as a Docker container, and authenticates using Grafana service account tokens. This implementation is particularly valuable for workflows requiring monitoring data analysis, metric exploration, or incident management directly within AI assistant conversations.
Nessuna recensione ancora. Sii il primo a recensire!
Accedi per unirti alla conversazione
Highlight AIList all teams.
List all users in an organization.
List all Grafana roles.
Get details for a Grafana role.
List assignments for a role.
List roles for users.
List roles for teams.
List permissions for a resource.
Describe a Grafana resource type.
Get the current Grafana identity — login, email, name, etc.
Search for dashboards by query, folder UID, tag, or starred.
Get a dashboard by UID, optionally a saved version.
List saved versions of a dashboard.
Update or create a new dashboard.
Get panel title, queries, datasource UID and type from a dashboard.
Execute one or more dashboard panel queries.
Extract specific parts of a dashboard using JSONPath expressions.
Get a compact summary of a dashboard without full JSON.
List datasources.
Get a datasource by UID or name.
Get example queries for a datasource type.
Execute a query against a Prometheus datasource.
List metric metadata.
List available metric names.
List label names matching a selector.
List values for a specific label.
Query and retrieve logs using LogQL.
List all available label names in logs.
List values for a specific log label.
Execute CloudWatch metric queries.
List alert groups from Grafana OnCall.
Get a specific alert group from Grafana OnCall by its ID.
Acknowledge, unacknowledge, resolve, or unresolve an alert group.
Retrieve a list of Sift investigations with an optional limit.
Retrieve an existing Sift investigation by its UUID.
Retrieve a specific analysis from a Sift investigation.
Detect elevated error patterns in Loki logs.
Detect slow requests from the relevant tempo datasources.
Fetch annotations with filters.
Create a new annotation.
Update specific fields of an annotation.
Delete an annotation by ID.
List dashboard snapshots with optional query and limit filters.
Get snapshot metadata and dashboard payload by snapshot key.
Create a dashboard snapshot from a full dashboard payload.
Delete a dashboard snapshot by snapshot key.
Render a stored dashboard or panel as a PNG image.
List provisioning repositories with their source URL, branch, sync state, and health.
Dry-run-apply a file from a provisioning repository and report admission validation errors.
