new_relic-automation

New Relic Automation — Automate New Relic Operations via Rube MCP

Skills Overview

Automate New Relic tasks through Composio’s Rube MCP server, including APM monitoring, alert configuration, dashboard management, NRQL queries, and infrastructure monitoring.

Use Cases

1. DevOps Pipeline Integration

Automatically trigger New Relic operations in CI/CD pipelines—for example, creating snapshots after deployment, validating performance metrics, or configuring environment monitoring. No need to log in manually to the New Relic console; with Rube MCP, you can complete all monitoring configuration in code.

2. Bulk Monitoring Resource Management

When you need to manage New Relic resources uniformly across multiple applications or environments, use this skill to create alert rules in bulk, synchronize dashboard configurations, or execute NRQL queries. It’s especially suitable for multi-environment deployments and standardized monitoring scenarios.

3. Automated Operations and Response

Build automated workflows so that when New Relic detects anomalies, it automatically runs diagnostic queries, collects relevant data, or triggers subsequent actions. Suitable for SRE teams building automated incident response and self-healing systems.

Core Capabilities

Tool Auto-Discovery and Execution

Use RUBE_SEARCH_TOOLS to dynamically query the currently available New Relic tool list, retrieving the latest tool schema and recommended execution steps. This avoids hard-coding tool names and keeps you synchronized with updates to tools on the Composio platform. Supports both single-step execution and multi-step complex workflows.

Connection Management and Status Validation

Use RUBE_MANAGE_CONNECTIONS to manage New Relic OAuth connections, automatically detecting connection status (ACTIVE/EXPIRED/INACTIVE). If a connection has expired, provides re-authentication guidance to ensure all operations run under a valid connection. Supports multi-account management and connection status monitoring.

NRQL Querying and Data Operations

Automate execution of NRQL (New Relic Query Language) queries to batch-extract APM, infrastructure, and log data. Supports pagination for large datasets, allowing query results to be passed to downstream tools or stored in Workbench for further analysis. Suitable for generating custom reports and data analysis tasks.

Common Questions

What is Rube MCP? How do I get started?

Rube MCP is an MCP (Model Context Protocol) server provided by Composio. It can be used without an API key. Simply add https://rube.app/mcp as the MCP server endpoint in your client configuration. After connecting successfully, you can use RUBE_SEARCH_TOOLS to verify service availability, and then use RUBE_MANAGE_CONNECTIONS to configure New Relic account authorization.

Why do I need to search tools before each execution?

The Composio tool list and schemas may be updated. Hard-coding tool names can cause calls to fail. RUBE_SEARCH_TOOLS returns not only the currently available tools, but also recommended execution steps, known issues, and input schemas. This is a key step to ensure the skill runs reliably long-term.

How do I handle API rate limits?

New Relic APIs have rate limits. For bulk operations, it’s recommended to:

Use ThreadPoolExecutor to execute independent operations in parallel to improve efficiency

When encountering rate limit errors, apply an exponential backoff retry strategy

For large bulk tasks, split them into multiple smaller batches and add delays

Use RUBE_REMOTE_WORKBENCH to run long-running bulk tasks remotely and avoid local timeouts

Does it support NRQL queries? How do I use it?

Fully supported. First, use RUBE_SEARCH_TOOLS to query NRQL-related tools (such as “execute NRQL query”), and then use RUBE_MULTI_EXECUTE_TOOL to execute the queries. Query results can be used to:

Generate custom reports

Trigger downstream automation actions

Store in Workbench for subsequent analysis

Validate post-deployment performance metrics

For large dataset queries, check pagination markers in the response (such as next_cursor) and continue fetching until all data is complete.