Grafana MCP: Production AI Observability on AWS

About this course

I have taught ~17000 students about monitoring with Prometheus and Grafana. Recently, I've been receiving several questions about how to deploy Grafana MCP in production. This course contains an extensive library of information to help you get up and running with MCP from the ground up. You will gain practical experience that you can immediately apply in your production systems.

At the end of the course, you'll have experience with the following:

• Running Grafana MCP locally with Docker and connecting it to a real Grafana instance.
• Connecting Grafana MCP to Grafana Cloud and query Cloud-backed metrics, logs, and dashboards.
• Deploying Prometheus, Loki, Grafana, and a demo app on AWS with Terraform.
• Deploying Grafana MCP on Amazon ECS Fargate and Amazon EKS with TLS, IAM, and Secrets Manager.
• Querying live dashboards, metrics, logs, and alerts from Cursor, Claude Code, Devin Desktop, and OpenCode.
• Explaining MCP architecture: clients, servers, tools, and transports (stdio, SSE, and streamable HTTP).
• Scoping Grafana service accounts and MCP tools so the assistant runs with least privilege, not admin.

Identifying MCP security risks and apply the controls you actually need on AWS.Choosing an ECS or EKS architecture for Grafana MCP and monitoring the MCP server itself.

Investigating incidents from your editor using live Prometheus and Loki data instead of screenshots.

Prices, discounts and availability are set by Udemy. We may earn a commission when you purchase through links on this site.