Deployment¶

Docker-based deployment with automated scripts. Supports standard, parallel, and incremental strategies.

Deployment Methods¶

Standard¶

Stops current container, starts new one. Causes brief downtime.

./fast-deploy.sh --migrate

Parallel (Reduced Downtime)¶

Starts new container alongside existing one, switches traffic when ready.

./fast-deploy.sh --parallel --migrate

Incremental (Faster)¶

Only updates what changed.

./fast-deploy.sh --incremental --migrate
# Skip resource copy (use cache):
./fast-deploy.sh --incremental --skip-resources --migrate

Database Handling¶

All deployments preserve the production database by default.

Run Migrations (preserve data)¶

./fast-deploy.sh --migrate

Backup¶

Automatic backup runs before every deployment.

# Manual backup:
./fast-deploy.sh --backup

# Restore from latest:
./fast-deploy.sh --restore

# Restore from specific file:
./fast-deploy.sh --restore /path/to/backup.sql

Reset Database¶

Danger

This deletes ALL production data. Requires explicit confirmation.

./fast-deploy.sh --reset-db

Health Checks & Rollback¶

After deployment, a health check verifies the app is running. Auto-rollback on failure.

./fast-deploy.sh --migrate                            # default: health check enabled
./fast-deploy.sh --migrate --skip-health-check        # skip (not recommended)
./fast-deploy.sh --migrate --health-endpoint /api/healthcheck

CI/CD (GitHub Actions)¶

Workflow in .github/workflows/deploy.yml:

Build & test on push to main
Auto-deploy to staging after successful build
Deploy to production on release or manual trigger
Auto-backup before deployment
Health check with auto-rollback

Manual Trigger¶

GitHub → Actions → "Deploy NFC Connect"
Click "Run workflow"
Choose environment, method, options

Deployment Files¶

File	Purpose
`deploy.py`	Main deployment script
`instance_up_with_resources.py`	Start/configure instances
`instance_down_with_resources.py`	Stop instances + resource backup
`fast-deploy.sh`	User-friendly shell wrapper

Troubleshooting¶

SSH issues: Check DigitalOcean access token, droplet name in properties.yml, SSH access for root user.

Migration failures: Check migration script, verify schema matches expected state, check error logs, restore from backup if needed. Never use --reset-db to fix migration failures without backing up first.

Health check failures: Check health endpoint, verify app logs, ensure container is running, check HTTP/HTTPS accessibility.