Documentation
Baudflow is a self-hosted network monitoring platform built with Phoenix LiveView. These docs cover everything from a one-command install to tuning adaptive schedules, breach policies, and scraping the Prometheus metrics endpoint.
Getting started
Overview
What baudflow is, the design principles behind it, and the full feature catalog: real-time visibility, health intelligence, alerting, and stack integration.
Read →Getting started
Get Baudflow running in under a minute: a Docker one-liner or a local Mix dev setup.
Read →Deployment
Run Baudflow in production: container image, environment, health, metrics, and backups.
Read →Guides
Self host
Deploy a self-hosted speed test with baudflow three ways: Docker run, Docker Compose, Kubernetes. Multi-arch non-root image, Ookla CLI bundled, no elevated capabilities, /health and /metrics included.
Read →Configuration
How settings, schedules, servers, thresholds, and notifications fit together, all from the Settings page.
Read →Schedules & adaptive cadence
Run multiple schedules at once, and let a breached schedule run faster until it recovers.
Read →Internet speed SLA
Internet speed SLA is the share of tests that hit your promised speed. How baudflow derives it on the fly, the auto/absolute/off threshold modes, breach streaks, and just-in-time health verdicts.
Read →Notifications
The four-layer alert pipeline (event, policy, template, channel), with ntfy and webhooks.
Read →Ping monitoring
A first-class ping console with a live per-sample diagnostics panel, using unprivileged TCP-connect RTT.
Read →Monitor ISP speed
Monitor ISP speed by building an auditable record of your own line with baudflow, then reading it out: scrape Prometheus /metrics into Grafana, embed the health heatmap, or screenshot the calendar for a provider.
Read →Reference
Configuration reference
Every Settings key, its default, and what it controls.
Read →Environment variables
Process env vars read at runtime: database, secrets, host, ports, and clustering.
Read →Prometheus metrics
The eight baudflow_* gauges exposed at GET /metrics, in text-exposition format.
Read →HTTP endpoints
The lightweight HTTP surface: /health, /metrics, and the embeddable heatmap.
Read →Comparison with speedtest-tracker
An honest, source-cited comparison of baudflow and speedtest-tracker — where each wins and where the gaps are, with links to the upstream issue tracker.
Read →Resources
Troubleshooting
Common issues: missing speedtest binary, database connection, ping reading 100% loss, and CI-vs-local test gates.
Read →Contributing
Local development: the justfile gates, the throwaway Docker test database, the fake CLI, and conventional-commit versioning.
Read →Changelog
Release history, generated from the project CHANGELOG.
Read →FAQ
Answers to the common questions about baudflow: hosting, data retention, databases, auth, licensing, and how it compares to speedtest-tracker.
Read →