Open source - Lumenstat

Everything that makes up Lumenstat is MIT licensed: the service code, the Umami patches, the tracker, the infrastructure code and these docs. Umami itself is MIT. A community operator can run the whole service from the repository without asking anyone for permission.

Repository layout

lumenstat

apps

packages

patches

deploy

infra

docs

scripts

examples

Path	What it holds
`apps/collector`	OpenTelemetry Collector configuration: tenant stamping, spanmetrics, dropping traces and logs
`apps/writer`	Stores API metrics, normalizes routes, serves the metrics API
`apps/dashboard`	The Next.js dashboard for projects and SCF operators
`packages/tracker`	Released tracker files, their integrity hashes, and the npm package
`patches/umami`	The changes to Umami v3.5.0, one commit per concern, applied by `scripts/sync-umami.sh`
`deploy`	Docker Compose and scripts for a single-server deploy
`infra`	OpenTofu for Cloudflare and Ansible for the host, for staging and production
`docs`	This site

Why a patch set instead of a fork

Umami is kept as an upstream release plus a short series of patches. Each patch does one thing, has its own switch, and leaves Umami behaving like upstream when the switch is off. That keeps every privacy and anti-inflation change reviewable on its own, and makes an Umami upgrade a matter of rebasing a few small commits instead of merging two diverged codebases.

Contributing

Bugs and requests: open an issue on GitHub, or post on the feedback board if you are not on GitHub.
Security problems: never in a public issue. Follow the security policy.
Pull requests: keep them small, run the tests for what you touched (npm test in apps/writer, pnpm vitest run src/lib/scf in vendor/umami, npm run build in apps/dashboard), and explain the reason for the change, not only what changed.
Docs: pages live in docs/. Preview with npx mint dev from that folder, and read docs/AGENTS.md for the writing rules.