Appsignal MCP

Appsignal MCP (Model Context Protocol)

A Model Context Protocol (MCP) server for Appsignal that allows AI assistants to fetch and analyze incident data.

Overview

This MCP server provides tools for AI assistants to:

1. Fetch incident details from Appsignal
2. Get incident samples
3. List recent incidents
4. Analyze incidents and suggest potential fixes

Prerequisites

• Node.js 14.x or higher
• An Appsignal account with API access
• An Appsignal Personal API Token

Installation

Option 1: Install from npm (Recommended)

You can install the package globally:

Or run it directly with npx:

Option 2: Install from Source

1. Clone this repository:
   git clone https://github.com/c20020207/appsignal-mcp.git cd appsignal-mcp
2. Install dependencies:
   npm install
3. Create a .env file with your Appsignal API token:
   APPSIGNAL_API_TOKEN=your_personal_api_token APPSIGNAL_APP_ID=your_default_app_id # Optional
   You can find your API token in your personal settings screen in Appsignal.

Usage

Running the MCP Server

Start the server:

For development with debug logging:

Integrating with AI Assistants

VS Code with Cursor

Add the following to your VS Code settings.json file:

If you installed from source, use this configuration instead:

Claude Desktop

Add the following to your Claude Desktop configuration:

If you installed from source, use this configuration instead:

Available Tools

getIncident

Get details about a specific incident.

Parameters:

• appId (string, required): Appsignal application ID
• incidentNumber (number, required): Incident number to fetch
• sampleId (string, optional): Specific sample ID to fetch

getIncidentSamples

Get samples related to an incident.

Parameters:

• appId (string, required): Appsignal application ID
• incidentNumber (number, required): Incident number
• limit (number, optional): Maximum number of samples to return (default: 10)

listIncidents

List recent incidents from an application.

Parameters:

• appId (string, required): Appsignal application ID
• limit (number, optional): Maximum number of incidents to return (default: 25)
• offset (number, optional): Offset for pagination (default: 0)
• state (string, optional): Filter by incident state ('open', 'closed', 'ignored')
• namespaces (array, optional): Filter by namespaces
• order (string, optional): Sort order ('last_occurrence', 'first_occurrence', 'count')

analyzeIncident

Analyze an incident and provide insights.

Parameters:

• appId (string, required): Appsignal application ID
• incidentNumber (number, required): Incident number to analyze

Example Prompts

Here are some example prompts you can use with AI assistants:

• "Fetch incident #123 from Appsignal app 'abc123'"
• "List the most recent open incidents from my Appsignal app"
• "Analyze incident #456 and suggest potential fixes"
• "Get samples for incident #789 to help debug the issue"

License

MIT