Deployment

This guide covers deploying group-ride on a Scaleway instance using Docker Compose, with automatic deployments triggered by a git tag.

Prerequisites

• A Scaleway account
• A GitHub Personal Access Token with read:packages scope (to pull the image from GHCR)
• Your bot configured — see Installation

---

Step 1 — Create a Scaleway instance

A STARDUST1-S (1 vCPU, 1 GB RAM, ~€1.80/month) is sufficient for this bot. A DEV1-S (2 vCPU, 2 GB RAM) gives more headroom if needed.

When creating the instance, select the Docker InstantApp image — Docker and Docker Compose are pre-installed and ready to use.

Once the instance is created, SSH in and verify:

ssh root@<YOUR_INSTANCE_IP>
docker --version
docker compose version

---

Step 2 — Authenticate to GitHub Container Registry

The Docker image is hosted on GHCR. You need a GitHub Personal Access Token (PAT) to pull it.

1. Go to GitHub → Settings → Developer settings → Personal access tokens → Tokens (classic)
2. Generate a token with the read:packages scope
3. On the instance, log in:

echo "<YOUR_PAT>" | docker login ghcr.io -u <YOUR_GITHUB_USERNAME> --password-stdin

---

Step 3 — Set up the deployment directory

mkdir -p /opt/group-ride && cd /opt/group-ride

Create the .env file:

cat > .env <<EOF
DISCORD_TOKEN=your_bot_token
DISCORD_CLIENT_ID=your_client_id
DISCORD_GUILD_ID=your_server_id
DISCORD_ANNOUNCEMENT_CHANNEL_ID=your_announcement_channel_id
DISCORD_FORUM_CHANNEL_ID=your_forum_channel_id
DATABASE_PATH=/app/data/group-ride.db
# DATABASE_URL=postgres://user:password@your-db-host:5432/group_ride
TZ=Europe/Paris
EOF

Keep this file private — it contains your bot token.

Download the Compose file:

curl -fsSL https://raw.githubusercontent.com/Slashgear/group-ride/main/docker-compose.yml -o docker-compose.yml

---

Step 4 — Start the bot

docker compose pull
docker compose up -d

Check that it is running:

docker compose ps
docker compose logs -f

You should see:

Group Ride bot is running

---

Step 5 — Enable automatic deployments (CI/CD)

The release workflow (tag → test → GitHub release → Docker image → deploy) is already set up. The deploy job SSHes into your instance and runs docker compose pull && docker compose up -d. You just need to give it the credentials.

5a — Generate a dedicated SSH key pair

On your local machine (not the instance):

ssh-keygen -t ed25519 -C "github-actions-deploy" -f ~/.ssh/group-ride-deploy -N ""

This creates two files:

• ~/.ssh/group-ride-deploy — private key (goes to GitHub)
• ~/.ssh/group-ride-deploy.pub — public key (goes to the instance)

5b — Authorise the key on the instance

ssh root@<YOUR_INSTANCE_IP> "echo '$(cat ~/.ssh/group-ride-deploy.pub)' >> ~/.ssh/authorized_keys"

5c — Add secrets to GitHub

Go to GitHub → your repo → Settings → Secrets and variables → Actions and add:

Secret name	Value
DEPLOY_HOST	Your instance IP address
DEPLOY_SSH_KEY	Contents of ~/.ssh/group-ride-deploy (the private key)

5d — Configure the GitHub environment

The deploy job uses the production environment, which lets you add protection rules (e.g. manual approval before each deploy).

Go to GitHub → Settings → Environments → production and configure as needed. If you skip this, the job still runs — the environment is optional.

---

Releasing a new version

Once everything is set up, releasing is a single command from your local machine:

# 1. Bump the version in package.json
#    e.g. change "version": "0.2.0" to "0.3.0"

# 2. Commit, then tag and push
bun run release

This triggers the full pipeline: tests → GitHub release → Docker image → SSH deploy.

---

Useful commands

# View live logs
docker compose logs -f

# Stop the bot
docker compose down

# Restart
docker compose restart

# Open a shell in the running container
docker compose exec bot sh

---

SQLite data

The database is stored in a named Docker volume (group-ride_data) and persists across container restarts and image updates. To back it up:

docker run --rm \
  -v group-ride_data:/data \
  -v $(pwd):/backup \
  busybox tar czf /backup/group-ride-backup-$(date +%Y%m%d).tar.gz /data

---

PostgreSQL data

If you set DATABASE_URL, the bot connects to your PostgreSQL instance instead of SQLite — no volume needed. Run the migrations once before starting the bot:

for f in src/adapters/database/postgres/migrations/*.sql; do
  psql "$DATABASE_URL" -f "$f"
done

For backup, use your provider’s snapshot feature or pg_dump:

pg_dump "$DATABASE_URL" > group-ride-backup-$(date +%Y%m%d).sql