logicmonitor

This MCP server provides a single, unified LogicMonitor assistant interface that exposes over 120 MCP tools for interacting with LogicMonitor. It supports multiple transport modes (STDIO for local usage, SSE/HTTP for remote access) and flexible authentication (no auth for development, bearer token, or OAuth/OIDC). By using the MCP prompts, you can perform read operations (like listing resources, status checks, and metric queries) as well as safe write operations when needed, with read-only mode enabled by default to protect data integrity. The server includes smart batching, rate limiting, and CSRF protection for session-based authentication, along with options to filter tools and customize logging formats for debugging.

To use the server, connect your MCP consumer (for example Claude, Cursor, Grafana Assistant, or VS Code integration) to the LogicMonitor MCP endpoint. You’ll authenticate with either a bearer token or via an OAuth flow, then select the LogicMonitor tools you want to run. Tool filtering lets you enable or disable specific capabilities, and the server will batch or throttle requests automatically to respect API limits. The included environment-based configuration supports quick local setup via a .env file or CLI flags, enabling teams to tailor the experience to their LogicMonitor environment.

Prerequisites:

• Node.js v18 or newer installed on the host
• npm or yarn package manager
• Access to the internet to install npm packages

Option A: Install and run with npm (recommended for development)

1. Ensure Node.js is installed: node -v and npm -v

2. Install the MCP server globally or run directly via npx:

   npx -y logicmonitor-mcp-server

3. If you want to customize environment variables, create a .env file or set them inline when starting the process, e.g. LM_COMPANY=your-subdomain LM_BEARER_TOKEN=your-token

Option B: Run via Docker (Docker must be installed)

1. Pull or run the container with environment variables configured:

   docker run -i --rm
   -e LM_COMPANY=your-subdomain
   -e LM_BEARER_TOKEN=your-token
   ghcr.io/monitoringartist/logicmonitor-mcp-server

2. Ensure the container is accessible to your MCP client via STDIO, SSE, or HTTP as configured by the server. You can also map ports if exposing HTTP/SSE externally.

Option C: Local development with Docker (VS Code installation prompts in docs show a Docker-based workflow)

1. Use the instructions in the project README to install via VS Code with the provided configuration UIs.
2. The Docker-based flow will pass LM_COMPANY and LM_BEARER_TOKEN into the container as shown in the examples.

Tips and common issues:

• Ensure your LogicMonitor company/subdomain (LM_COMPANY) and API Bearer Token (LM_BEARER_TOKEN) are valid; invalid credentials will fail authentication.
• If running locally, enable read-only mode by default if you want to prevent modification actions during exploration.
• Use the .env file approach to simplify repeated deployments; the server supports CLI flags, environment variables, or .env configuration.
• If you enable TLS/HTTPS, provide proper certificates and configure your transport layer accordingly.
• When using Docker, remember to set appropriate resource limits to handle the 125 MCP tools without exhausting memory or CPU.
• For debugging, switch log format to JSON for structured logs or human-readable for quick inspection; both are supported.