New Relic MCP Server

A Model Context Protocol (MCP) server that provides New Relic observability tools for integration with Cursor and other MCP-compatible AI assistants.

Features

This MCP server provides comprehensive New Relic tools organized into the following categories:

Discovery (Entity and Account Management)

• convert_time_period_to_epoch_ms - Convert time periods to epoch milliseconds
• get_entity - Fetch entities by GUID or search by name
• list_related_entities - List related entities (1 hop away)
• list_available_new_relic_accounts - List all available accounts
• list_dashboards - List dashboards for an account
• get_dashboard - Get dashboard details
• list_entity_types - List all entity types
• search_entity_with_tag - Search entities by tag

Data Access

• execute_nrql_query - Execute NRQL queries against NRDB
• natural_language_to_nrql_query - Convert natural language to NRQL and execute

Alerting

• list_alert_conditions - List alert conditions for a policy
• search_incident - Search alert incidents
• list_alert_policies - List alert policies
• list_recent_issues - List open issues
• list_synthetic_monitors - List synthetic monitors

Incident Response

• analyze_deployment_impact - Analyze deployment performance impact
• generate_alert_insights_report - Generate alert analysis report
• generate_user_impact_report - Generate user impact report
• get_entity_error_groups - Get error groups from Errors Inbox
• list_change_events - List deployment and config changes

Performance Analytics

• analyze_entity_logs - Analyze logs for patterns and anomalies
• analyze_golden_metrics - Analyze throughput, response time, error rate
• analyze_kafka_metrics - Analyze Kafka consumer lag, throughput
• analyze_threads - Analyze JVM thread metrics
• analyze_transactions - Identify slow and error-prone transactions
• list_garbage_collection_metrics - Get GC and memory metrics
• list_recent_logs - List recent log entries

Advanced Analysis

• comprehensive_entity_analysis - Full entity health analysis
• correlate_metrics_and_events - Correlate metrics with events
• analyze_service_dependencies - Analyze service dependencies
• generate_slo_report - Generate SLO compliance report

Prerequisites

• Python 3.10+ (required by MCP SDK)
• New Relic account with API access
• New Relic User API Key

Installation

1. Clone the repository:

   cd /Users/vickykumar/projects/newrelic-mcp
   

2. Create and activate a virtual environment (Python 3.10+ required):

   python3.11 -m venv venv  # or python3.10, python3.12, etc.
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Configure environment variables:

   cp env.example .env
   

   Edit .env and set your New Relic credentials:

   NEW_RELIC_API_KEY=NRAK-your-api-key-here
   NEW_RELIC_ACCOUNT_ID=your-account-id
   

Getting Your New Relic API Key

1. Go to New Relic API Keys
2. Click "Create a key"
3. Select "User" as the key type
4. Give it a name (e.g., "MCP Server")
5. Copy the key (starts with NRAK-)

Usage with Cursor

Option 1: Add to Cursor MCP Settings (Recommended)

1. Open Cursor Settings (⌘/Ctrl + ,)
2. Search for "MCP" or navigate to Features → MCP Servers
3. Click "Add MCP Server" or edit your ~/.cursor/mcp.json
4. Add the following configuration:

{
  "mcpServers": {
    "newrelic": {
      "command": "/Users/vickykumar/projects/newrelic-mcp/venv/bin/python",
      "args": ["/Users/vickykumar/projects/newrelic-mcp/server.py"],
      "env": {
        "NEW_RELIC_API_KEY": "NRAK-your-api-key",
        "NEW_RELIC_ACCOUNT_ID": "your-account-id",
        "NEW_RELIC_REGION": "US"
      }
    }
  }
}

Note: Update the paths if your project is located elsewhere.

Option 2: Using the startup script

Create a wrapper script for easier configuration:

#!/bin/bash
# save as: run-newrelic-mcp.sh

export NEW_RELIC_API_KEY="NRAK-your-api-key"
export NEW_RELIC_ACCOUNT_ID="your-account-id"

cd /Users/vickykumar/projects/newrelic-mcp
./venv/bin/python server.py

Then configure Cursor to use this script:

{
  "mcpServers": {
    "newrelic": {
      "command": "/Users/vickykumar/projects/newrelic-mcp/run-newrelic-mcp.sh",
      "args": []
    }
  }
}

Example Queries in Cursor

Once configured, you can ask the AI assistant to:

Discovery

• "List all my New Relic accounts"
• "Find all APM applications with the tag environment=production"
• "Show me the entities related to my payment service"

Data Access

• "Execute this NRQL query: SELECT count(*) FROM Transaction SINCE 1 hour ago"
• "Show me errors from the last 30 minutes"
• "What's the throughput trend for my API?"

Alerting

• "List all open incidents"
• "Show me alert policies for my account"
• "What synthetic monitors are failing?"

Incident Response

• "Analyze the impact of the last deployment on my checkout service"
• "Generate an alert insights report for issue ID xyz123"
• "What error groups are affecting my service?"

Performance Analytics

• "Analyze the golden metrics for entity ABC123"
• "Show me slow transactions in my payment service"
• "What are the error patterns in my application logs?"

Advanced Analysis

• "Run a comprehensive health analysis on my API service"
• "Generate an SLO report for my checkout service with 99.9% target"
• "Correlate metrics with recent deployments"

Testing the Server

You can test the server manually:

# Activate virtual environment
source venv/bin/activate

# Set environment variables
export NEW_RELIC_API_KEY="NRAK-your-key"
export NEW_RELIC_ACCOUNT_ID="your-account-id"

# Run the server (it will wait for MCP protocol messages on stdin)
python server.py

Troubleshooting

Server not starting

• Check that Python 3.9+ is installed
• Verify the virtual environment is activated
• Ensure all dependencies are installed

Authentication errors

• Verify your API key is correct and starts with NRAK-
• Check that the account ID is correct
• Ensure the API key has the necessary permissions

No data returned

• Verify the entity GUID is correct
• Check the time window (some queries default to 60 minutes)
• Ensure your account has data for the requested time period

Cursor not finding the server

• Check the path in your MCP configuration
• Ensure the server script is executable
• Look at Cursor logs for MCP-related errors

Environment Variables

Variable	Required	Default	Description
NEW_RELIC_API_KEY	Yes	-	Your New Relic User API Key
NEW_RELIC_ACCOUNT_ID	Yes	-	Default New Relic Account ID
NEW_RELIC_REGION	No	US	Region: US or EU
MAX_LOOKBACK_MINUTES	No	10080	Max lookback period (7 days)
MAX_RESULT_LIMIT	No	2000	Max results per query
DEFAULT_LIMIT	No	100	Default result limit

Architecture

newrelic-mcp/
├── server.py              # Main MCP server entry point
├── requirements.txt       # Python dependencies
├── .env                   # Environment configuration
├── .env.example          # Example configuration
└── app/
    ├── __init__.py
    ├── config.py          # Configuration management
    ├── nr_client.py       # New Relic GraphQL client
    └── tools/
        ├── __init__.py
        ├── discovery.py       # Entity/account tools
        ├── data_access.py     # NRQL query tools
        ├── alerting.py        # Alert/incident tools
        ├── incident_response.py   # Incident analysis
        ├── performance_analytics.py  # Performance tools
        └── advanced_analysis.py   # Advanced analysis

License

MIT License - See LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.