Connect Grafana Cloud to Claude Code
Search, read and write Grafana Cloud from Claude Code. Arlet links the two through one MCP connection and takes care of keys and authentication.
Arlet makes Grafana's official MCP server available from the AI client you already use. Sign in to Grafana Cloud and choose read, query or write access to complete the connection.
The AI can search and read dashboards, query Prometheus, Loki, Tempo and other data sources, manage alerts, incidents and on-call, and run Assistant investigations. This covers Grafana Cloud stacks, and the available tools are those provided by Grafana's official server.
How to connect Grafana Cloud to Claude Code
- 1
In the Arlet dashboard, open Integrations β Add an integration and choose Grafana Cloud. Sign in with Link Grafana Cloud, name the connection and click Create.
- 2
In your terminal, run claude mcp add --transport http arlet https://mcp.arlet.dev/mcp/ to add Arlet as an MCP server.
- 3
Start Claude Code, run /mcp, select arlet and click Authenticate.
- 4
Your browser opens. Sign in with Google and click Allow on the consent screen. The connection is complete.
The connection is per user and works across every project on the same machine.
What you can do
Grafana Cloud tools that Claude Code can call.
search_dashboardsSearch dashboards.
get_dashboard_summaryGet a dashboard summary.
get_dashboard_panel_queriesGet the queries behind panels.
update_dashboardCreate or update a dashboard.
list_datasourcesList data sources.
query_prometheusQuery Prometheus metrics.
query_loki_logsQuery Loki logs.
query_loki_patternsExtract log patterns.
query_pyroscopeQuery Pyroscope profiles.
alerting_manage_rulesManage alert rules.
list_incidentsList incidents.
create_incidentCreate an incident.
get_current_oncall_usersGet who is on call now.
create_annotationAdd an annotation.
create_investigationStart an Assistant investigation.
ask_assistantAsk Grafana Assistant a question.
The tool list follows what Grafana Cloud's official MCP server provides and may change.
Example uses
Read a dashboard
Read a dashboard's panels and queries and explain the current state in words.
Correlate logs and metrics
Confirm an error-rate spike in Prometheus and pull the matching Loki logs to narrow the cause.
Handle on-call
Check who is on call, open an incident and leave annotations.
Tune alerts
List existing alert rules and adjust thresholds or add new rules.
Prompts you can use as-is
- Read the API Overview dashboard and summarize the current state
- Find services whose error rate rose in the last 30 minutes in Prometheus and pull the matching error logs from Loki
- Who is on call right now? Open an incident for the checkout outage
- List the alert rules for the checkout service and add a p95 latency alert
