New Relic Integrates Kiro and Commits to MCP for AI-Assisted Development

New Relic has long been the standard for observability across engineering teams, but the rise of agent-powered IDEs has forced a rethinking of where application state context actually lives. This week, according to HPCwire, the company has announced its integration with Kiro, the agent-oriented IDE that AWS unveiled just weeks ago. The technical link enabling this connection is MCP, Anthropic's Model Context Protocol.

This is no minor move: New Relic manages telemetry data from millions of services running in production. Making that data directly accessible from the environment where developers write and debug code fundamentally changes the feedback loop.

What the integration delivers in practice

Kiro is an IDE with built-in agents capable of executing development tasks autonomously or with assistance: code generation, review, and deployment of changes. The integration with New Relic via MCP allows those agents to query metrics, traces, and production alerts without leaving the development environment.

The specific workflow described in the announcement looks something like this:

• An agent in Kiro detects a failure in a branch before merging.
• It queries the New Relic MCP server to retrieve the error history of that service in production.
• It proposes a fix grounded in real data, not assumptions.

For teams already using New Relic as their observability layer, this eliminates the manual context switch between the monitoring platform and the IDE. It's a concrete use case for which MCP was designed: exposing external system data in standardized form so any compatible agent can consume it.

Why MCP instead of a proprietary integration

This is perhaps the most significant aspect of the news for those following the Claude ecosystem. New Relic could have built a Kiro-specific plugin or a direct integration with AWS APIs. Instead, it chose to implement an MCP server.

The reasoning is pragmatic: an MCP server is natively compatible with Claude Code, with any MCP client, and by extension with any IDE or agent that adopts the protocol. New Relic is not betting only on Kiro; it is betting on the standard. This means the same integration should work in Claude Code if your team configures the server in your `claude_desktop_config.json` or registers it as a tool in your agent environment.

The pattern we are seeing in 2026 is consistent: data and infrastructure platforms that want to be accessible to AI agents do not build N bilateral integrations; they build an MCP server and delegate compatibility to the protocol.

Who benefits from this

The New Relic + Kiro integration via MCP has direct utility for specific profiles:

• Platform engineering teams that already have New Relic deployed and want their developers working with production context in the IDE without switching tools.
• Developers using Claude Code who want to connect their agents to real observability data. If New Relic publishes the MCP server publicly and with clear documentation, adoption in the Claude ecosystem will be immediate.
• Teams evaluating Kiro as an alternative to other agent-powered IDEs: the availability of mature MCP integrations is an increasingly common selection criterion.

What remains unclear is whether New Relic's MCP server will be open access or restricted to customers on certain plans. That detail significantly affects the true scope of the announcement.

Our take

The news confirms something we have observed for months: MCP is moving beyond being "Anthropic's protocol" to become foundational infrastructure for the broader agent ecosystem. That a company the size of New Relic adopts it as a primary integration vector is a clear signal of consolidation, not experimentation.