Open-source Datadog MCP alternatives
Datadog's MCP server is open source, so you can read how search_datadog_logs and get_datadog_trace reach your account before connecting it. The servers below publish their code too. For observability that sees your logs, traces, and incident data, reading the source tells you exactly which queries an agent can run and what it can change.
These are the open-source observability servers people weigh against Datadog. Several run locally, a few are hosted endpoints, and all let you pin the version you trust and audit the repo first.
The 8 best open-source alternatives
Open-core and broad: Grafana Labs' official server queries dashboards, Prometheus, Loki, incidents, alerts, and OnCall, the widest open overlap with Datadog's logs-metrics-alerting span.
Set up Grafana →Error tracking with the source open: Sentry's official server pulls issues, stack traces, and events and runs Seer root-cause analysis, the focused choice for debugging application errors.
Set up Sentry →A small, inspectable metrics server: this community Prometheus build runs PromQL instant and range queries, discovers metrics, and inspects scrape targets, easy to read end to end before you trust it.
Set up Prometheus →OpenTelemetry-native and open: SigNoz's official server gives an agent traces, logs, metrics, dashboards, and alerts, matching Datadog's breadth on an OTel foundation you can audit.
Set up SigNoz →Incident response, source published: PagerDuty's official server exposes incidents, services, schedules, and orchestrations across 64 tools, read-only by default, for the on-call side of an outage.
Set up PagerDuty →- HoneycombOfficial
High-cardinality debugging with open code: Honeycomb's official server queries traces, metrics, and logs, runs BubbleUp, and manages Boards, Triggers, and SLOs over a hosted endpoint.
Set up Honeycomb → - New RelicOfficial
Full-stack monitoring you can read: New Relic's official server runs NRQL, searches entities, analyzes golden metrics, and triages incidents, a hosted but open alternative across the same surface as Datadog.
Set up New Relic → Adjacent rather than a monitor: AWS Labs' open-source server runs any AWS CLI command with validation and a read-only mode, useful for pulling CloudWatch or service state when your telemetry lives in AWS.
Set up AWS (AWS Labs) →
How to choose
Every option ships its code, so decide by surface. Grafana and SigNoz are the broadest open overlaps with Datadog, SigNoz on OpenTelemetry. New Relic and Honeycomb cover full-stack and high-cardinality work respectively. Sentry is for errors, PagerDuty for incidents, Prometheus for metrics alone, and the AWS server is the way in when your observability data sits inside AWS itself. Read the repo before granting query access.
FAQ
- Is the Datadog MCP server open source?
- Yes. Datadog publishes the source for its official server, which includes log, metric, trace, and incident tools. Every alternative on this page is open source as well, so you can read each before connecting it to your telemetry.
- Which open-source alternatives can I also self-host?
- Grafana, Prometheus, SigNoz, and the AWS server ship local builds, so the process and credentials stay on your infrastructure. Sentry and PagerDuty offer self-hosted servers too. Honeycomb and New Relic publish their source but run as hosted endpoints.