n9e

This MCP server integrates Nightingale with MCP clients, allowing AI assistants to manage alerts, targets, incidences, and collaboration data via natural language. Once configured, the MCP server exposes a suite of toolsets that the client can invoke to query active and historical alerts, list and inspect alert rules, browse monitored targets, manage mutes and notification rules, handle event pipelines, and query users or business groups. The included toolsets can be selectively enabled via the N9E_TOOLSETS environment variable or the --toolsets flag to minimize context usage and focus the AI on the tasks you need.

To use, configure an MCP client (such as Cursor) to point at this server. The client will expose commands like list_active_alerts, list_targets, list_alert_rules, create_mute, and list_event_pipelines, among others. When you send natural language prompts, the agent translates them into tool invocations against Nightingale data through this MCP server, enabling workflows like querying current alerts, reviewing alert rules for a business group, or inspecting the status of monitored targets. Be aware of authentication and base URL requirements: ensure N9E_TOKEN and N9E_BASE_URL are properly set so the client can authenticate and reach the Nightingale API.

Prerequisites:

• Go 1.21+ for building Nightingale MCP Server or access to the prebuilt MCP package on npm/npx as configured.
• Node.js and npm (if using npx invocation via npm registry).
• Access credentials for your Nightingale instance (API token).

Install and run (example using npx as described in the repo):

# Prerequisite: ensure you have an API token from Nightingale
# Configure environment variables as needed (see below)
export N9E_TOKEN=your-api-token
export N9E_BASE_URL=http://your-n9e-server:17000

# Start the MCP server via npx (as per README configuration)
npx -y @n9e/n9e-mcp-server stdio

If you prefer to embed the configuration into a file (recommended for repeatable deployments):

{
  "mcpServers": {
    "nightingale": {
      "command": "npx",
      "args": ["-y", "@n9e/n9e-mcp-server", "stdio"],
      "env": {
        "N9E_TOKEN": "your-api-token",
        "N9E_BASE_URL": "http://your-n9e-server:17000",
        "N9E_TOOLSETS": "alerts,targets,datasource,mutes,busi_groups,notify_rules,alert_subscribes,event_pipelines,users"
      }
    }
  }
}

For Docker-based deployment (if supported in your environment), you would typically pull the image and run with appropriate environment variables, but this repository demonstrates using npx for simplicity and alignment with MCP client tooling.

Tips:

• Start with a minimal toolset (e.g., alerts and targets) using N9E_TOOLSETS to reduce context size and token usage. You can expand later as needed.
• Ensure the Nightingale API token has the necessary permissions to read alerts, rules, and user data.
• If you encounter authentication errors, verify N9E_BASE_URL is reachable from the host running the MCP server and that the token is valid.
• The client configuration example uses a single server named nightingale; you can rename or extend the mcpServers map to support multiple Nightingale instances.
• When updating toolsets, you may need to restart clients to pick up changes.
• Check Nightingale v8.0.0+ compatibility and ensure the MCP server version matches the Nightingale API version you target.
• If you experience slow prompts or token expiry, consider rotating tokens and refreshing configurations in your MCP client.