New Relic MCP

New Relic MCP connects an agent to a New Relic account so it can query the observability data an engineering team already collects: application traces, host and infrastructure metrics, browser and mobile telemetry, logs, and the alerting layer built on top of them. Access is bound to the API key or authenticated profile used to connect, so the agent sees only the accounts and entities that identity is entitled to under the platform's role-based access controls. The central capability is querying. The agent can run NRQL against the telemetry database directly, or describe what it wants in plain language and have the server produce and execute the query for it, which removes the main barrier for people who do not write NRQL fluently. Around that sit entity operations for finding services, hosts, and dashboards by name or tag, plus alerting operations for listing policies and their conditions and pulling the incident history behind a page. It is aimed at engineers and on-call responders who want investigation to happen where the code is. During an incident, the assistant can locate the affected service, chart its error rate and latency, read the matching log lines, and check which alert condition fired, all without a context switch to the browser. The same access supports routine work such as capacity reviews, dashboard summaries, and pre-deploy health checks.