Grafana MCP
Lets AI assistants query Grafana Cloud metrics, logs, and traces and manage dashboards and alerts.
Publisher
Grafana Labs
Product Details
The Grafana Cloud hosted MCP server gives AI agents secure, authenticated access to a full observability stack through the Model Context Protocol, with no local server to run. SRE, platform, and on-call teams can query Prometheus/Mimir, Loki, Tempo, and Pyroscope with PromQL, LogQL, and TraceQL, and list, read, and update dashboards and data sources.
Agents can manage alert rules, create and update incidents in Grafana Cloud IRM, view on-call schedules, and generate deep links back to the Grafana Cloud UI. Typical uses include asking why latency is spiking, summarizing active incidents, building visualizations, and auditing dashboards and alerts. Permissions follow Grafana Cloud access policies and the same RBAC model as the rest of the stack.
Key Use Cases
Incident Root Cause Triage and Diagnosis
SRE and on-call engineers execute PromQL and LogQL queries, inspect Sift investigations, and correlate error logs with infrastructure metrics during high-severity outages.
Observability as Code and Alert Auditing
DevOps teams programmatically discover underutilized dashboards, inspect alert rule configurations via alerting_manage_rules, and generate targeted deep links for team postmortems.
Explore detailed deployment path
Requires Gemini. Access integration prerequisites, specialized agent configuration guides, and implementation documentation.