New Relic MCP Server for Claude

Content

New Relic MCP Server is New Relic's official Model Context Protocol bridge for giving AI development tools controlled access to New Relic observability context. It is useful when Claude, Codex, VS Code, Windsurf, Gemini CLI, or another MCP-capable client needs to inspect live telemetry, alerts, incidents, entities, dashboards, logs, or NRQL-backed analysis during debugging and operations work.

The fit is strongest for read-only investigation: entity discovery, alert triage, recent issue review, performance analysis, deployment impact review, and natural-language queries over New Relic data. Keep permissions narrow, use include-tags to limit the available tool categories, and require a human to review any query, report, or operational conclusion before it changes an incident response plan.

Features

• Official New Relic AI MCP Server documented by New Relic.
• Public preview setup flow with prerequisites, organizational role requirements, and preview enablement in New Relic.
• Remote HTTP endpoint at https://mcp.newrelic.com/mcp/.
• OAuth and user API key authentication options.
• Claude Code setup with OAuth or API key authentication.
• Codex CLI setup with OAuth command-line registration or config-based API key headers.
• Claude Desktop setup through mcp-remote when OAuth is needed.
• VS Code workspace setup through .vscode/mcp.json.
• Tool filtering with the include-tags HTTP header.
• Tool categories for discovery, data access, alerting, incident response, performance analytics, and advanced analysis.
• New Relic RBAC enforcement based on the configured user API key or OAuth profile.

Use Cases

• Ask Claude to list available New Relic accounts, entities, dashboards, and related entities before an incident review.
• Pull recent alerts, open issues, alert policies, and synthetic monitor context for a service without leaving the coding assistant session.
• Convert a plain-English performance question into an NRQL-backed analysis, then have an operator review the query and result before acting on it.
• Analyze recent logs, golden metrics, transactions, Kafka metrics, garbage collection metrics, or thread metrics while debugging an application issue.
• Review deployment impact and change events for an affected New Relic entity during release triage.
• Generate a draft alert insight or user-impact summary for an incident handoff while preserving a clear source trail back to New Relic data.

Installation

Claude Code

For OAuth, add the remote MCP endpoint:

claude mcp add newrelic --transport http https://mcp.newrelic.com/mcp/

Start Claude Code, run /mcp, select the New Relic server, and complete the browser-based OAuth flow.

For API key authentication, use a New Relic user API key and keep the real key out of committed configuration:

claude mcp add newrelic https://mcp.newrelic.com/mcp/ --transport http --header "Api-Key: NRAK-YOUR-KEY-HERE"

Codex CLI

New Relic documents a config-based API key path for Codex:

[mcp_servers.new-relic]
url = "https://mcp.newrelic.com/mcp/"
env_http_headers = { "api-key" = "NEW_RELIC_API_KEY" }

For OAuth, New Relic documents the Codex command-line setup:

codex mcp add new-relic-mcp-server --url "https://mcp.newrelic.com/mcp/"

Then start Codex, run /mcp, and complete the authentication flow.

Claude Desktop

Claude Desktop can use mcp-remote for the New Relic remote MCP endpoint:

{
  "mcpServers": {
    "new-relic-mcp": {
      "command": "npx",
      "args": ["mcp-remote", "https://mcp.newrelic.com/mcp/"]
    }
  }
}

VS Code

Use a workspace .vscode/mcp.json file:

{
  "servers": {
    "new-relic-mcp": {
      "url": "https://mcp.newrelic.com/mcp/",
      "type": "http"
    }
  }
}

Add headers only when the workspace is private and the credential handling is approved.

Tool Scope

Use include-tags to narrow the tools exposed to the model:

{
  "servers": {
    "new-relic-mcp-server": {
      "url": "https://mcp.newrelic.com/mcp/",
      "type": "http",
      "headers": {
        "include-tags": "discovery,alerting"
      }
    }
  }
}

Useful starting scopes:

• discovery: account, dashboard, entity, relationship, and entity-type discovery.
• data-access: NRQL execution and natural-language-to-NRQL query workflows.
• alerting: alert conditions, alert policies, recent issues, incidents, and synthetic monitors.
• incident-response: deployment impact, alert insight reports, user-impact reports, error groups, and change events.
• performance-analytics: logs, golden metrics, Kafka metrics, threads, transactions, garbage collection metrics, recent logs, and performance risk groups.
• advanced-analysis: combined analysis workflows such as deployment impact, log analysis, alert insights, user impact, and natural-language NRQL.

Examples

Incident triage

Use New Relic MCP to list open issues, recent alerts, related entities, recent change events, and the most relevant dashboards for this service. Do not change anything.

NRQL review

Draft the NRQL query needed to compare request latency and error rate for the last 60 minutes. Show the query before executing it.

Deployment impact

Analyze deployment impact for this entity GUID over the last two hours and summarize evidence, time windows, and confidence. Keep this read-only.

Tool inventory

Before calling any New Relic tool, list which New Relic MCP tags and tool categories are available in this session.

Source Notes

• New Relic's MCP overview describes the New Relic AI MCP Server as a centralized bridge for connecting supported AI development tools to New Relic observability context.
• The overview documents the preview status, supported clients, prerequisites, required New Relic organizational roles, authentication options, and public preview enablement.
• The setup guide documents Claude Code, Codex CLI, Claude Desktop, VS Code, Windsurf, and Gemini CLI setup paths.
• The tool reference documents natural-language use, include-tags, RBAC enforcement, and tool categories for discovery, data access, alerting, incident response, performance analytics, and advanced analysis.
• The documented MCP endpoint is https://mcp.newrelic.com/mcp/.

Duplicate Check

Before drafting this entry, the current upstream content tree and live open PRs were checked for New Relic, newrelic, new-relic-mcp-server, mcp.newrelic.com, docs.newrelic.com/docs/agentic-ai/mcp, and related title variants. Existing generic SRE and MCP setup entries only mention New Relic as an example observability product. Existing Datadog, Grafana, and Sentry MCP entries cover different observability providers. No dedicated New Relic MCP Server entry, New Relic MCP source URL duplicate, or open duplicate PR was found.

Editorial Disclosure

Submitted as a community MCP entry by oktofeesh1. This listing is based on New Relic's official MCP overview, setup guide, and tool reference, with no paid placement or affiliate relationship.

• Source repositorygithub.com 2026-07-19T11:20:19-07:00
• Documentationdocs.newrelic.com
• Websitenewrelic.com
• Submitted by oktofeesh12026-06-04