Manager installation

The Fleetdock Manager is the control plane. This page covers installing it, the environment variables it reads, and how to run it safely in production.

System requirements

You also need Docker Engine 24+ and Docker Compose v2.23+, with ports 80 and 443 free. TLS is handled for you by the bundled Caddy, which obtains a Let's Encrypt certificate automatically — a domain pointed at the host is optional but recommended.

Database requirements

The Manager stores metadata in PostgreSQL. The Compose stack bundles a local Postgres container for development. In production, point FLEETDOCK_DATABASE_URL at managed PostgreSQL (Neon, RDS, Cloud SQL, or any PostgreSQL 14+) with sslmode=require.

Install

curl -sSL https://fleetdock.dev/install.sh | sh -s -- --domain db.example.com

This installs Docker if missing, writes /opt/fleetdock/{docker-compose.yml,.env} with generated secrets, starts the stack, and prints the dashboard URL and bootstrap password. Re-running it upgrades in place without touching .env.

Manage it afterwards with the fleetdock command it installs — status, logs, update, domain, gateway, doctor, backup-config, uninstall.

From source

git clone https://github.com/Fleetdock/fleetdock.git
cd fleetdock
cp .env.example .env
./scripts/generate-secrets.sh >> .env

docker compose -f docker-compose.yml -f docker-compose.build.yml up --build -d

Environment variables

Variables use the FLEETDOCK_* prefix. The most important ones:

See .env.example in the repository for the full list, including optional SMTP variables for email notifications and gateway configuration for external database access.

HTTPS and reverse proxy

The stack ships Caddy on 80/443 with automatic certificates, so there is nothing to set up. Nothing else is published — the control plane's 8080 stays on the internal network.

If you replace it with your own proxy, there is no path routing to reproduce: the control plane serves the API and the dashboard on one port, so a single upstream is the whole config.

db.example.com {
    encode zstd gzip
    reverse_proxy fleetdock:8080
}

Persistent storage

The bundled Postgres writes to a Docker named volume (postgres_data). In production, your managed PostgreSQL provider is the source of truth — the metadata database holds users, server enrollment, encrypted credentials, backup destinations, and job history. Losing it without a backup means re-enrolling every server and re-entering every external credential.

Verify

curl -fsS https://db.example.com/healthz    # {"status":"ok"}
curl -fsS https://db.example.com/readyz     # {"status":"ready"} once the DB is up

Then sign in and change the bootstrap admin password on the Profile page. Interactive API docs are also available on this site under API reference (with an interactive playground). The Manager itself serves Redoc at /docs over the embedded OpenAPI spec.

Update

fleetdock update              # latest
fleetdock update --tag v0.2.0 # a specific release

Or, from a source checkout:

git fetch origin && git checkout v0.2.0
docker compose -f docker-compose.yml -f docker-compose.build.yml up --build -d

Migrations run automatically on API start. Confirm GET /readyz returns ready, then spot -check login and a connection test. Watch the Operations page for stuck jobs after an upgrade.

Uninstall

docker compose down          # stop containers, keep the database volume
docker compose down -v       # also delete volumes — destroys the metadata database