Do I need Docker?

No. There isn’t even a Dockerfile in the repository.

Feasible is a single self-contained Go binary. Copy it to a server, run ./feasible serve, and the whole product is up - the dashboard, the ingest endpoint, the settings screens, the docs, the background jobs, all in one process.

What you also don’t need:

• No Postgres, no MySQL. Storage is SQLite, and the driver is pure Go, so there’s no C library to link and no client to install.
• No ClickHouse or other column store. Reports read from pre-built roll-ups in the same database file.
• No Redis, no Kafka, no queue. Background jobs are rows in SQLite.
• No reverse proxy required. Put one in front if you want TLS termination the usual way, but nothing breaks without it.
• No Node at runtime. The dashboard, the CSS and the tracker are compiled at build time and embedded in the binary.

So the deployment is: a binary, a data directory, and a systemd unit. Backups are file copies, and there’s a db backup command that writes a consistent snapshot of every database while it’s running.

You can absolutely put it in a container if that’s how your infrastructure works - nothing stops you, and a FROM scratch image with one static binary in it is about as small as containers get. It just isn’t a requirement, and we’re not going to publish a compose file with four services in it to make a pageview counter look serious.

Everything you need is 1 core and 512 MB, and the three commands to start are on can I run Feasible myself.