# New Relic MCP for AI Agents AI Agent Connect

> New Relic MCP lets you query your entire infrastructure and application stack using natural language. It connects your AI agent directly to the NerdGraph API so you can track entities, run NRQL queries, and monitor active alerts without leaving your workspace.

## Overview
- **Category:** loved-by-devs
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_Rk3fKfy2EBc5U3UTb8bGKnyj2pxBgzOIOazg0CbK/ai-agent-connect
- **Tags:** apm, infrastructure-monitoring, nrql, log-management, real-time-alerts, graphql

## Description

You can stop jumping between browser tabs just to check your system's health. By connecting your New Relic account to your agent, you get a direct line to your observability data. You can ask for a summary of your application's performance, check if a specific host is struggling, or pull up the latest alert details in seconds. It handles the heavy lifting of the NerdGraph API, letting you focus on fixing problems rather than navigating complex menus. You can even run custom NRQL queries to dig into specific datasets or check your SLOs to see if you're meeting your performance targets. This Connector is part of the Vinkius catalog, making it easy to plug into your existing workflow and get immediate answers about your stack's uptime and performance. Think about the last time you had to debug a production incident. You probably had to open a dozen tabs, find the right dashboard, and run several manual queries just to see the big picture. This Connector changes that experience. You can now stay in your primary workspace and just talk to your agent. It can pull up your golden metrics, list your active alerts, and even give you the configuration for your existing dashboards. If you need to go deeper, you can ask it to run a specific NRQL query to find exactly where the bottleneck is. Whether you are an SRE looking for a smoking gun or a developer trying to verify a deployment, you get the information you need without the friction of manual navigation. It's about getting to the root cause faster.

## Tools

### get_dashboard
Retrieves the specific configuration details for a New Relic dashboard. Use this to understand how a specific view is built.

### list_dashboards
Lists all the dashboards available in your New Relic account. This makes it easy to find existing monitoring views.

### list_entities
Finds and lists monitored entities like hosts, containers, and browser apps. It helps you identify specific assets in your stack.

### get_apm_summary
Pulls golden metrics like error rates and response times for your applications. This helps you see high-level health at a glance.

### get_entity_details
Pulls metadata for a specific monitored entity like a host or container. This provides granular details on individual components.

### get_me
Shows the current user information and the accounts you have access to. Use this to verify your active session.

### list_accounts
Lists all the New Relic accounts your current user can access. This is helpful for managing multiple environments.

### list_alerts
Shows active issues and alerts across your monitored infrastructure. This allows you to quickly see what needs attention.

### list_service_levels
Lists your Service Level Indicators (SLOs) to track performance goals. Use this to monitor your uptime targets.

### run_nrql
Executes raw NRQL queries to get specific datasets or custom metrics. This is your go-to tool for deep data dives.

## Prompt Examples

**Prompt:** 
```
What's the current error rate for our production web app?
```

**Response:** 
```
### Application Health Summary
**Application:** Web-Frontend-Prod

| Metric | Value |
| :--- | :--- |
| **Error Rate** | 0.05% |
| **Apdex Score** | 0.98 |
| **Avg Response Time** | 156ms |

*Status: Normal operating range.*
```

**Prompt:** 
```
List all the active alerts in my account right now.
```

**Response:** 
```
### Active Alerts
I found 3 active issues:

*   **High CPU Usage** on host `prod-db-01` (Critical)
*   **500 Errors** on `Auth-API` (Warning)
*   **Latency Spike** on `Inventory-Service` (Warning)

Would you like me to pull more details on any of these?
```

**Prompt:** 
```
Run this NRQL: SELECT average(duration) FROM Transaction FACET appName SINCE 1 DAY AGO
```

**Response:** 
```
### Average Duration per App (Last 24h)

| App Name | Avg Duration |
| :--- | :--- |
| Web-Frontend | 0.12s |
| Auth-API | 0.08s |
| DB-Proxy | 0.45s |

*Results formatted as a table.*
```

## Capabilities

### Run custom NRQL queries
Execute raw queries to pull specific datasets and custom metrics from your account.

### Fetch golden metrics
Get real-time summaries for error rates, Apdex scores, and response times.

### Monitor active alerts
Retrieve a list of current issues and active alerts across your infrastructure.

### Retrieve dashboard configs
Pull configuration details for your custom observability dashboards.

### Track performance SLOs
Check your Service Level Indicators to see if you're meeting performance goals.

### Discover monitored entities
List and search for APM applications, hosts, containers, and browser apps.

## Use Cases

### Rapid Incident Response
An SRE asks the agent to list_alerts to see what is firing, then uses run_nrql to find the specific error spike in the logs.

### Post-Deployment Verification
A developer asks for the get_apm_summary for a new service to ensure the error rate hasn't jumped since the latest push.

### Reliability Reporting
A product manager asks the agent to check list_service_levels to see if the application is meeting its 99.9% uptime goal.

### Infrastructure Auditing
An admin uses list_entities to find all active containers in a specific region to check for over-provisioning or orphaned resources.

## Benefits

- Run NRQL queries without leaving your workspace to get immediate data on your specific datasets.
- Get instant summaries of golden metrics like Apdex scores and error rates using get_apm_summary.
- Stay on top of production issues by listing active alerts with list_alerts to see what is firing right now.
- Quickly find specific infrastructure components using list_entities and get_entity_details to identify bottlenecks.
- Track your performance targets easily by pulling SLO data with list_service_levels to ensure reliability.
- Access dashboard configurations quickly using list_dashboards to understand your current monitoring setup.

## How It Works

The bottom line is you get a natural language interface for your entire New Relic observability stack.

1. Subscribe to the New Relic MCP through the Vinkius marketplace.
2. Enter your New Relic User API Key in the configuration settings.
3. Ask your AI client questions about your app's health or infrastructure status.

## Frequently Asked Questions

**Can I use the New Relic MCP to check my app's health?**
Yes. You can ask your AI agent for a summary of your application's performance or check specific metrics like error rates and response times instantly.

**Does the New Relic MCP support custom NRQL queries?**
It does. You can provide raw NRQL strings to your agent to pull specific datasets, custom metrics, or complex filtered results.

**How do I connect New Relic to my AI agent?**
First, subscribe to the Connector on Vinkius. Then, provide your New Relic User API Key in the configuration to link your account.

**Can I see my SLOs using the New Relic MCP?**
Yes. You can ask the agent to list your Service Level Indicators (SLOs) to see if your applications are meeting their performance targets.

**Can the New Relic MCP find my active alerts?**
Yes. You can ask the agent to list active issues and alerts across your accounts to see what needs immediate attention.

**Is the New Relic MCP good for SRE teams?**
It is perfect for SREs who need to quickly identify issues, check host statuses, and run queries without navigating complex dashboards.

**Where do I find my User API Key?**
In New Relic, go to the account dropdown > API keys. Look for a key of type 'USER'. This is required for NerdGraph access.

**Can I run any NRQL query through this server?**
Yes! Use the `run_nrql` tool and provide your Account ID and the query string. This allows for extremely flexible data retrieval.

**What are APM golden metrics?**
Golden metrics are the most critical performance indicators for an application: Response Time, Throughput, Error Rate, and Apdex (user satisfaction).