SSM Runtime Configuration

This is a reference pattern. Learn from the approach, adapt to your context — don't copy verbatim.

Audience: Service developers (Lambda, ECS, frontend). If you're using CDK to deploy infrastructure, see CDK Bootstrap Configuration.

Problem: Configuration scattered across .env files, CI/CD secrets, and cloud resources becomes out of sync and hard to manage.

Solution: Use AWS SSM Parameter Store as single source of truth for runtime configuration, with minimal .env files only for local development overrides.

---

Pattern

Architecture:

Infrastructure Deployment → Creates/Updates SSM Parameters
                                      ↓
                            SSM Parameter Store
                            (Single Source of Truth)
                                      ↓
                    ┌─────────────────┴─────────────────┐
                    ↓                                   ↓
            Local Development                    CI/CD Pipeline
         (with .env.local overrides)          (SSM values only)

Key Components:

• SSM Parameter Store: Runtime configuration for all environments
• infrastructure/.env: Bootstrap config (AWS account, region) - deployment only
• service/.env.local: Local development overrides - optional, not committed

---

Why This Pattern?

Benefits:

• Single Source of Truth: All runtime config in SSM, no sync issues
• Environment Isolation: Separate parameters per environment
• No Secrets in Repo: Sensitive values stay in AWS
• CI/CD Compatible: Pipelines fetch from SSM, no local files needed
• Runtime Flexibility: Change config without redeployment

Use Cases:

• AWS serverless applications
• Multi-environment deployments
• Projects with sensitive configuration
• CI/CD pipelines

---

Implementation

SSM Namespace Structure:

/{project-id}/{environment}/{service}/VARIABLE_NAME

Examples:
/my-app/local/frontend/API_ENDPOINT
/my-app/dev/frontend/API_ENDPOINT
/my-app/prod/backend/DATABASE_URL

Bootstrap Configuration (infrastructure/.env):

# Minimal config for CDK deployment
CDK_DEFAULT_ACCOUNT=123456789012
CDK_DEFAULT_REGION=eu-central-1

Local Overrides (frontend/.env.local):

# Override SSM values for local development only
API_ENDPOINT=http://localhost:3000

Retrieving Parameters:

// At application startup
import { SSMClient, GetParameterCommand } from '@aws-sdk/client-ssm';

const ssmClient = new SSMClient({ region: process.env.AWS_REGION });

async function getConfig(environment: string, service: string) {
  const path = `/${PROJECT_ID}/${environment}/${service}/`;
  // Fetch all parameters under path
  // Apply .env.local overrides if environment === 'local'
  return config;
}

---

File Strategy

What to Commit:

• ✅ infrastructure/.env - Bootstrap config (no secrets)
• ✅ .env_TEMPLATE files - Documentation
• ❌ .env.local - Local overrides (gitignored)
• ❌ .env.dev, .env.prod - Runtime config (in SSM instead)

Template Files:

# .env_TEMPLATE
# Purpose: Document expected variables
# Instructions: Copy to .env.local and customize

# API Endpoint (from SSM by default)
# API_ENDPOINT=http://localhost:3000

---

CLI Integration

# Get single parameter
ssm-params get /{project-id}/dev/frontend/API_ENDPOINT

# Export all parameters for service
ssm-params export-service dev frontend

# Output as .env format
ssm-params export-service dev frontend > .env

---

Tradeoffs

Pros:

• Single source of truth
• No sync issues
• Secure secret storage

Cons:

• Requires AWS access for local development
• Slightly slower startup (network call to SSM)
• More complex than simple .env files

When NOT to Use:

• Non-AWS projects
• Simple applications with no sensitive config
• Projects without multiple environments

---

Related Patterns

• CDK Bootstrap Configuration - For CDK projects: how to write outputs to SSM without causing context caching issues
• Environment Validation - Validate config early
• Environment Deployment Strategy - How environments are deployed
• Security Baseline - Secret management principles

---

Progressive Improvement

If the developer corrects a behavior that this skill should have prevented, suggest a specific amendment to this skill to prevent the same correction in the future.