Skip to main content

Cloud Database Deployment

Run Cascadia with a managed cloud database (AWS RDS, Google Cloud SQL, Azure Database).

Architecture​

┌─────────────────────────────────────┐
│        Application Server(s)        │
│                                     │
│  ┌───────────────────────────────┐  │
│  │        Cascadia App           │  │
│  │   (Core + Vault + Jobs)       │  │
│  └───────────────┬───────────────┘  │
└──────────────────┼──────────────────┘
                   │ TLS
                   │
┌──────────────────▼──────────────────┐
│         Cloud Provider              │
│                                     │
│  ┌───────────────────────────────┐  │
│  │     Managed PostgreSQL        │  │
│  │                               │  │
│  │  - AWS RDS                    │  │
│  │  - Google Cloud SQL           │  │
│  │  - Azure Database             │  │
│  │                               │  │
│  │  Features:                    │  │
│  │  - Automated backups          │  │
│  │  - Multi-AZ failover          │  │
│  │  - Point-in-time recovery     │  │
│  │  - Automatic updates          │  │
│  └───────────────────────────────┘  │
└─────────────────────────────────────┘

When to Use​

  • Production environments
  • Compliance requirements for managed backups
  • Need for automated failover
  • Prefer managed database operations
  • Budget available for managed services

Cloud Provider Setup​

AWS RDS​

  1. Create RDS Instance

    Engine: PostgreSQL 18+
    Instance: db.t3.medium (adjust for workload)
    Storage: 100GB gp3
    Multi-AZ: Yes (production)

  2. Security Group

    • Allow inbound 5432 from application servers only

  3. Connection String

    DATABASE_URL=postgresql://cascadia:PASSWORD@mydb.abc123.us-east-1.rds.amazonaws.com:5432/cascadia?sslmode=require

Google Cloud SQL​

  1. Create Instance

    Engine: PostgreSQL 18
    Machine: db-custom-4-16384 (4 vCPU, 16GB)
    Storage: 100GB SSD
    High Availability: Regional

  2. Authorized Networks

    • Add application server IPs or use Cloud SQL Proxy

  3. Connection String (via proxy)

    DATABASE_URL=postgresql://cascadia:PASSWORD@127.0.0.1:5432/cascadia

    Or with Unix socket:

    DATABASE_URL=postgresql://cascadia:PASSWORD@/cascadia?host=/cloudsql/project:region:instance

Azure Database for PostgreSQL​

  1. Create Server

    Tier: General Purpose
    vCores: 4
    Storage: 100GB
    High Availability: Zone Redundant

  2. Firewall Rules

    • Add application server IPs

  3. Connection String

    DATABASE_URL=postgresql://cascadia@server:PASSWORD@server.postgres.database.azure.com:5432/cascadia?sslmode=require

    Note: Azure uses username@server format.

Deployment​

Step 1: Create Cloud Database​

Follow your cloud provider's instructions above.

Step 2: Create Database Schema​

# From your local machine with DATABASE_URL set
cd CascadiaApp
npx drizzle-kit push

Step 3: Deploy Application​

# On application server
cd deployments/cloud-database
cp .env.example .env
# Edit .env with your cloud DATABASE_URL

docker-compose up -d

Cost Optimization​

AWS RDS​

  • Use Reserved Instances for predictable workloads
  • Consider Aurora Serverless for variable workloads
  • Use read replicas for read-heavy applications

Google Cloud SQL​

  • Use committed use discounts
  • Consider Cloud SQL Insights for query optimization
  • Scale down during off-hours (if applicable)

Azure​

  • Use Reserved Capacity
  • Consider Hyperscale for very large databases
  • Use Flexible Server for dev/test (cost-effective)

Security Best Practices​

  1. Never expose database publicly

    • Use VPC peering or private endpoints

  2. Use SSL/TLS

    • Always include sslmode=require in connection string

  3. Rotate credentials regularly

    • Use AWS Secrets Manager, GCP Secret Manager, or Azure Key Vault

  4. Enable audit logging

    • Track database access for compliance

  5. Encrypt at rest

    • Enabled by default on all major providers

Backup and Recovery​

AWS RDS​

# Automated backups: Configure retention period (1-35 days)
# Point-in-time recovery: Within backup retention window
# Manual snapshots: Keep before major changes

Google Cloud SQL​

# Automated backups: Daily, 7-day retention default
# Point-in-time recovery: Up to 7 days
# On-demand backups: Before migrations

Azure Database​

# Automated backups: 7-35 days retention
# Geo-redundant backups: For disaster recovery
# Long-term retention: Up to 10 years

Next Steps​