Open-source New Relic MCP alternatives
New Relic publishes its MCP server's source, so you can read how it runs NRQL and resolves entities before wiring it into a client. Even so, it only runs as New Relic's hosted endpoint. Every option below is open source as well, which lets you audit the code, pin a version, and in several cases run the server against your own stack.
The list spans broad observability platforms and focused tools for metrics, errors, and on-call. What they share is a public repo you can clone and read end to end before granting access.
The 8 best open-source alternatives
Grafana Labs' server is open source and queries dashboards, Prometheus, Loki, incidents, alerts, and OnCall. You can read exactly which data sources it touches and run it against a Grafana stack you operate.
Set up Grafana →Sentry's official server is open source and focused on errors: pull issues, stack traces, and events, and run Seer root-cause analysis, with source you can vet before it reads your project data.
Set up Sentry →Open source and metrics-only, the Prometheus server runs PromQL instant and range queries, discovers metrics, and inspects scrape targets, a small repo you can audit and point at your own Prometheus.
Set up Prometheus →OpenTelemetry-native and open source, the SigNoz server exposes traces, logs, metrics, dashboards, and alerts. The whole stack is yours to read and run, a broad observability alternative without a closed endpoint.
Set up SigNoz →On-call is the focus of the open-source PagerDuty server: incidents, services, schedules, teams, and orchestrations, read-only by default. Reading the repo confirms exactly what an agent can trigger before you connect it.
Set up PagerDuty →- DatadogOfficial
Datadog's remote server is open source and the broadest peer here: search logs, query metrics, pull APM traces, inspect monitors, and investigate incidents, with code you can audit even though the backend is Datadog's.
Set up Datadog → - HoneycombOfficial
Query-centric and open source, the Honeycomb server runs queries and BubbleUp over traces, metrics, and logs and manages Boards, Triggers, and SLOs. The repo shows how it drives those investigations.
Set up Honeycomb → Adjacent rather than a direct swap: AWS Labs' open-source server runs any AWS CLI command, which reaches CloudWatch and other monitoring services if your telemetry lives in AWS. Useful when observability is one part of a wider AWS footprint.
Set up AWS (AWS Labs) →
How to choose
Among the open-source options, Grafana and SigNoz are the broadest and can run on your own infrastructure, while Datadog's server is open even though its backend is not. Sentry, Prometheus, PagerDuty, and Honeycomb each go deep on errors, metrics, on-call, or query-first investigation. AWS is the outlier, reaching monitoring only as part of its CLI. Whichever you choose, the open-source benefit is the same: read the repo before granting it access to your telemetry.
FAQ
- Is the New Relic MCP server open source?
- Yes, the source is published, so you can audit how it runs NRQL and resolves entities. It still only runs as New Relic's hosted endpoint, though. Every alternative on this page is open source too, and several can run against your own stack.
- Which open-source observability server can I fully self-host?
- Grafana, Prometheus, Sentry, and SigNoz can run locally against telemetry you control. Datadog's server is open source but talks to Datadog's hosted backend, so reading the code is possible while the data still leaves your network.