Deploy

This part has one guide for each place to run the stack, and each guide ends with pbmb admin status passing against the new address. It is for the person who chooses where the four containers run: the api, the worker, Cerbos and Postgres.

Every guide runs the release image ghcr.io/enum-studio/powerbi-metrics-book:<version> pinned by its digest, with pbmb from the powerbi-metrics-book package on PyPI at the same version. Each fetches its files from the commit the release’s tag names.

Pick one and follow it from start to end:

Guide	What runs where	Choose it when
One VM with Docker Compose	The compose file pbmb init writes, with Caddy in front for TLS, and nightly backups encrypted with age to two buckets off the VM.	You can run a Linux VM. This is the reference deployment, and the cheapest: we recommend a VM from Hetzner or Netcup, which cost much less than Azure, GCP or AWS.
Dokploy or Coolify on a VM	The same compose file as a Dokploy compose application, with Dokploy’s Traefik giving the domain and certificate. Backups as on the VM.	You already run Dokploy or Coolify and want its dashboard.
Railway	One Railway project from one file, .railway/railway.ts: a service per container, the identities as sealed variables, and Railway’s volume backups for Postgres.	You want no server to look after.
Azure Container Apps	One Bicep template: three container apps, PostgreSQL Flexible Server reachable only from the virtual network, Key Vault for every value, and a registry for the Cerbos image.	Your company already runs Power BI in Azure.

There is no Kubernetes chart yet; that page lists what the containers already meet.

Each guide says who acts at each step: you, at the terminal, or a person with a right you may not have, such as editing DNS or signing in to a dashboard. pbmb login in every guide needs a person to enter a code in a browser.

Run the quickstart first to see the stack work on one machine. Then come back here for the address people will open.