Metricbarn stores what your systems emit and makes it answerable in one query language — so a spike on a graph is two clicks away from the trace, the log line and the person who should hear about it.
# install the collector and point it at your workspace curl -sSL https://cdn.metricbarn.org/install.sh | sh barn auth login --workspace acme-prod barn agent start --autodiscover # services found: 14 · golden signals streaming in 9s
Build and share views with a PromQL-style query bar, then pin them to a wall board your whole team watches.
dash.metricbarn.orgThreshold, anomaly and burn-rate monitors with rotations, escalation policies and quiet hours that hold.
alerts.metricbarn.orgPull metrics and billing signals from your cloud accounts without running anything of your own.
cloud.metricbarn.orgEverything the UI does is an endpoint: write points, run queries, manage monitors from your own tooling.
api.metricbarn.orgInstrumentation guides, the full query reference and SDK snippets for Go, Python, Node and Rust.
docs.metricbarn.orgLive health for ingest, query, alerting and the web app, with 90 days of incident history in the open.
status.metricbarn.orgOne line per host or container. It autodiscovers your services and starts shipping golden signals immediately.
Latency, traffic, errors and saturation appear on generated dashboards — nothing to build before you get value.
Wrap any query in a monitor, choose a rotation, and let burn-rate rules keep the pages honest.
Free up to 10 hosts and 30 days of retention. No credit card, no sales call.
How we stopped scanning cold blocks for high-cardinality label filters, and what it cost us in write amplification.
A practical guide to error budgets: which windows to pick, and why the two-window rule stops the 3am false page.
Forty-one minutes of delayed points, the queue metric we were not watching, and the four changes that shipped after.
Rolling 30-day availability. Subscribe to updates on the status page — incidents are posted there before they reach support.
