tg-deployment
Production deployment patterns for the World of Darkness Django application. Use when deploying to staging/production, configuring security settings, setting up Redis cache, planning database migrations, or preparing rollback procedures. Triggers on deployment tasks, production configuration, security hardening, or environment-specific settings.
When & Why to Use This Skill
This Claude skill provides comprehensive production deployment patterns for Django applications, focusing on security hardening, performance optimization, and reliable release management. It offers structured guidance for environment configuration, Redis caching, database migrations, and disaster recovery procedures to ensure stable and secure production environments.
Use Cases
- Production Environment Setup: Configuring environment-specific settings and secure variables for Django applications to ensure consistency across staging and production.
- Security Hardening: Implementing SSL redirection, HSTS, and secure cookie policies to meet industry-standard production security requirements.
- Cache & Session Management: Setting up Redis as a high-performance cache backend and session store to improve application scalability and response times.
- Safe Database Migrations: Executing complex migrations with integrated backups, data integrity checks, and maintenance mode handling to prevent data loss.
- Incident Recovery: Utilizing documented rollback procedures for both code and database to quickly restore service in the event of a failed deployment.
| name | tg-deployment |
|---|---|
| description | Production deployment patterns for the World of Darkness Django application. Use when deploying to staging/production, configuring security settings, setting up Redis cache, planning database migrations, or preparing rollback procedures. Triggers on deployment tasks, production configuration, security hardening, or environment-specific settings. |
Deployment Patterns
Environment Configuration
Settings Structure
tg/settings/
├── __init__.py # Auto-loads based on DJANGO_ENVIRONMENT
├── base.py # Common settings (all environments)
├── development.py # DEBUG=True, local cache
└── production.py # Full security, Redis, HTTPS
Set DJANGO_ENVIRONMENT=production in .env for production.
Required Production Environment Variables
# Critical - must be set
DJANGO_ENVIRONMENT=production
SECRET_KEY=<generated-secret-key>
DJANGO_ALLOWED_HOSTS=example.com,www.example.com
CSRF_TRUSTED_ORIGINS=https://example.com
# Database (PostgreSQL recommended)
DB_NAME=tg_db
DB_USER=tg_user
DB_PASSWORD=<secure-password>
DB_HOST=localhost
DB_PORT=5432
# Redis cache
REDIS_URL=redis://127.0.0.1:6379/1
Generate secret key: python -c "from django.core.management.utils import get_random_secret_key; print(get_random_secret_key())"
Security Configuration
Production Security Settings (in production.py)
# HTTPS enforcement
SECURE_SSL_REDIRECT = True
SESSION_COOKIE_SECURE = True
CSRF_COOKIE_SECURE = True
# HSTS (1 year)
SECURE_HSTS_SECONDS = 31536000
SECURE_HSTS_INCLUDE_SUBDOMAINS = True
SECURE_HSTS_PRELOAD = True
# Content security
SECURE_CONTENT_TYPE_NOSNIFF = True
SECURE_BROWSER_XSS_FILTER = True
X_FRAME_OPTIONS = "DENY"
SECURE_REFERRER_POLICY = "same-origin"
# Cookie security
SESSION_COOKIE_HTTPONLY = True
SESSION_COOKIE_SAMESITE = "Lax"
CSRF_COOKIE_SAMESITE = "Strict"
Verify with: python manage.py check --deploy
Redis Cache Setup
Installation
pip install django-redis redis
Configuration
CACHES = {
"default": {
"BACKEND": "django_redis.cache.RedisCache",
"LOCATION": os.environ.get("REDIS_URL", "redis://127.0.0.1:6379/1"),
"OPTIONS": {
"CLIENT_CLASS": "django_redis.client.DefaultClient",
"CONNECTION_POOL_KWARGS": {"max_connections": 50},
"RETRY_ON_TIMEOUT": True,
"IGNORE_EXCEPTIONS": True,
"COMPRESSOR": "django_redis.compressors.zlib.ZlibCompressor",
},
}
}
# Session storage in Redis
SESSION_ENGINE = "django.contrib.sessions.backends.cache"
SESSION_CACHE_ALIAS = "default"
Verification
from django.core.cache import cache
cache.set('test', 'value', 60)
cache.get('test') # Should return 'value'
Database Migration Deployment
Pre-deployment
# Backup database
pg_dump production_db > backup_$(date +%Y%m%d_%H%M%S).sql
# Validate existing data
python manage.py validate_data_integrity --verbose
python manage.py validate_data_integrity --fix # If issues found
Deployment Steps
# 1. Enable maintenance mode
touch /var/www/tg/maintenance.flag
# 2. Pull code
git fetch origin && git checkout <branch> && git pull
# 3. Install dependencies
pip install -r requirements.txt --upgrade
# 4. Run migrations
python manage.py showmigrations # Preview
python manage.py migrate
# 5. Collect static
python manage.py collectstatic --noinput
# 6. Restart services
sudo systemctl restart gunicorn
sudo systemctl restart nginx
# 7. Disable maintenance mode
rm /var/www/tg/maintenance.flag
Post-deployment
# Health check
curl -I https://your-domain.com/
python manage.py monitor_validation
Rollback Procedures
Code-only Rollback
git log --oneline # Find previous commit
git checkout <previous-commit>
sudo systemctl restart gunicorn
Full Rollback (Code + Database)
# 1. Stop application
sudo systemctl stop gunicorn
# 2. Restore database
dropdb production_db
createdb production_db
pg_restore -d production_db backup_TIMESTAMP.sql
# 3. Revert code
git checkout <previous-commit>
python manage.py migrate <app> <previous-migration>
# 4. Restart
sudo systemctl start gunicorn
Deployment Checklist
Pre-deployment
- All tests passing locally
- Database backup created and verified
- Rollback plan documented
- Team notified of deployment window
During deployment
- Maintenance mode enabled
- Code pulled and dependencies updated
- Data validation passed
- Migrations applied successfully
- Static files collected
- Services restarted
Post-deployment
- Health check passed
- Key functionality tested (login, create, edit)
- Error logs clean
- Maintenance mode disabled
- Monitor for 30 minutes
Monitoring Commands
# Validation health check
python manage.py monitor_validation
# JSON output for monitoring tools
python manage.py monitor_validation --json
# Analyze specific time period
python manage.py monitor_validation --period 48