New Relic MCP Integration

Weldable's New Relic MCP integration gives your AI agents access to New Relic's observability platform for querying metrics, managing alerts, and retrieving application performance data through natural language. New Relic was named a Leader in the 2026 IDC MarketScape for Worldwide AIOps and has launched an SRE Agent that provides automated incident triage and root cause analysis. Weldable connects New Relic to your other tools so your agents can turn observability data into action across your incident response and project management workflows.

Your agent handles requests in plain English. Ask it to check error rates for a service, pull transaction traces, or list active alert violations, and Weldable maps the intent to the correct New Relic API call.

Use cases

Performance regression detection

After a deployment, your agent queries New Relic for response time and error rate metrics for the affected application. It compares the current values against a baseline from the previous 24 hours and flags any regressions that exceed a defined threshold. If a regression is detected, the agent posts a warning to Slack with the specific metrics and creates a high-priority issue in Linear or Jira. Teams catch performance problems early instead of waiting for customer reports.

Automated alert triage

Your agent checks New Relic for active alert violations and groups them by policy and condition. It pulls the relevant metric data for each violation, determines whether the issue is ongoing or recovering, and posts a structured summary to the on-call channel in Slack. Violations that correlate with recent deployments get flagged with the release version and commit details. This gives responders immediate context instead of raw alert notifications.

Weekly application health digest

Your agent queries New Relic for key application metrics on a weekly cadence: average response time, throughput, error percentage, and Apdex score. It compiles these into a formatted report, highlights any week-over-week changes, and posts it to Slack or Google Docs. Engineering leadership gets a consistent view of application health without logging into New Relic to build custom reports.

Infrastructure capacity planning

Your agent pulls host-level metrics from New Relic: CPU usage, memory consumption, disk utilization, and network throughput across your fleet. It identifies hosts that are consistently running above 80% capacity and flags them for scaling. The report goes to the infrastructure team with specific recommendations. This turns reactive scaling (responding to outages) into proactive planning.

SLA compliance tracking

Your agent queries New Relic for uptime and response time data, calculates SLA metrics against your committed targets, and generates a monthly compliance report. If any service is trending toward an SLA breach, the agent alerts the team early with the specific metrics that are at risk. This automates the reporting that otherwise requires manual data exports and spreadsheet calculations.

How it works

Connect your New Relic account by providing API keys. Weldable stores them securely and handles authentication for every request. Your agent gets access to APM data, infrastructure metrics, alerts, and NRQL queries based on the key permissions.

Describe what you need in natural language. Weldable maps your intent to the correct New Relic API endpoint or translates it into a NRQL query. Your agent can chain New Relic data with other services: detect a performance issue, create a ticket, and page the on-call engineer in one flow.

Tips

NRQL is your most powerful query tool. New Relic Query Language lets your agent run precise queries against any data in your New Relic account. Saying "average response time for the checkout service in the last 6 hours" translates to a NRQL query that returns exactly what you need. The more specific your request, the more targeted the query.

Application names must match exactly. New Relic identifies applications by their configured name. Make sure your agent references the exact application name as it appears in New Relic. Mismatches return empty results without an obvious error.

Alert policies group related conditions. New Relic organizes alert conditions under policies. When your agent queries violations, it can filter by policy name to focus on the relevant area, such as "infrastructure" or "web transactions." This prevents alert noise from unrelated systems.

Time windows change the story. The same metric can look very different depending on the time range. A 5-minute window shows spikes, while a 24-hour window shows trends. Be explicit about time ranges in your queries to get the right perspective for your use case.

Facet queries break data down by dimension. NRQL's FACET clause lets your agent group results by attributes like host, transaction name, or environment. This is useful for identifying which specific component is responsible for an aggregate metric change.