Dockerize App

Create a tested docker-compose.yml that spins up the full application stack — services, databases, seed data — so any developer or agent can run the entire dev environment with a single command.

Workflow

Copy this checklist and track progress:

Task Progress:
- [ ] Step 1: Analyze the codebase
- [ ] Step 2: Create Dockerfiles
- [ ] Step 3: Create docker-compose.yml
- [ ] Step 4: Set up data seeding
- [ ] Step 5: Test the full stack
- [ ] Step 6: Update AGENTS.md

Step 1: Analyze the Codebase

Explore the repository to determine:

• Services: What application services exist? (API server, frontend, workers, etc.)
• Languages/runtimes: Node, Python, Go, Ruby, etc. — drives Dockerfile base images.
• Package managers: npm, yarn, pnpm, pip, poetry, go mod, etc.
• Databases: PostgreSQL, MySQL, MongoDB, Redis, etc. Check for ORMs, migration files, connection strings, or .env references.
• Existing Docker artifacts: Any Dockerfile, docker-compose.yml, or .dockerignore already present.
• Environment variables: Check .env.example, .env.sample, config files, or code references to process.env / os.environ.
• Seed data: Look for seed scripts, fixtures, SQL dumps, or migration files that populate initial data.
• Inter-service communication: How do services talk to each other? (HTTP, gRPC, message queues, shared DB)

If the project has an existing docker-compose.yml, evaluate whether it needs updating rather than replacing it.

Step 2: Create Dockerfiles

For each application service that needs one, create a Dockerfile in that service's root directory.

Principles:

• Use specific version tags for base images (e.g., node:20-slim, not node:latest)
• Use multi-stage builds when a build step is needed (compiled languages, frontend asset builds)
• Copy dependency manifests first, install, then copy source — maximizes layer caching
• Run as a non-root user in the final stage
• Use .dockerignore to exclude node_modules, .git, .env, etc.

If a Dockerfile already exists, leave it as-is unless it's broken or clearly outdated.

Create a .dockerignore in each service root if one doesn't exist:

node_modules
.git
.env
.env.*
*.log
dist
build
.cache

Adapt the ignore list to the language/framework.

Step 3: Create docker-compose.yml

Create docker-compose.yml at the repository root.

Requirements:

• Every application service gets its own services: entry, built from its Dockerfile
• Databases and infrastructure (Postgres, Redis, etc.) use official images with pinned versions
• Use named volumes for database persistence
• Use a shared network so services can communicate by service name
• Map host ports for local access (avoid conflicts with common local services)
• Use env_file pointing to .env or inline environment vars with sensible defaults
• Set depends_on with condition: service_healthy where health checks are defined
• Add health checks for databases and critical services

Template structure:

services:
  db:
    image: postgres:16
    environment:
      POSTGRES_USER: ${DB_USER:-app}
      POSTGRES_PASSWORD: ${DB_PASSWORD:-password}
      POSTGRES_DB: ${DB_NAME:-app_dev}
    volumes:
      - db_data:/var/lib/postgresql/data
    ports:
      - "${DB_PORT:-5432}:5432"
    healthcheck:
      test: ["CMD-SHELL", "pg_isready -U ${DB_USER:-app}"]
      interval: 5s
      timeout: 3s
      retries: 5

  app:
    build:
      context: .
      dockerfile: Dockerfile
    ports:
      - "${APP_PORT:-3000}:3000"
    env_file: .env
    depends_on:
      db:
        condition: service_healthy
    volumes:
      - .:/app
      - /app/node_modules

volumes:
  db_data:

Adapt this to the actual project — the above is just a starting point for a typical Node + Postgres setup.

Step 4: Set Up Data Seeding

The goal: after docker-compose up, the app is ready to use with realistic sample data.

Strategy (pick the best fit):

1. ORM seeds/migrations exist → Run them automatically. Add a seed service or an entrypoint script that waits for the DB, runs migrations, then seeds.

2. SQL dump available → Mount it into the DB container's init directory. For Postgres: mount .sql files into /docker-entrypoint-initdb.d/. For MySQL: mount into /docker-entrypoint-initdb.d/.

3. No seed data exists → Create a minimal seed script. Write a script that inserts reasonable sample data. Place it in scripts/seed.sh or scripts/seed.sql.

Entrypoint pattern for auto-seeding:

Create a scripts/docker-entrypoint.sh for application services that need to run migrations + seeds on startup:

#!/bin/sh
set -e

echo "Waiting for database..."
# Use wait-for-it, dockerize, or a simple loop
until nc -z db 5432; do sleep 1; done

echo "Running migrations..."
# npm run migrate / python manage.py migrate / etc.

echo "Seeding database..."
# npm run seed / python manage.py seed / etc.

echo "Starting application..."
exec "$@"

Make sure the entrypoint is executable and referenced in the Dockerfile or docker-compose entrypoint.

Step 5: Test the Full Stack

Run the stack and verify everything works:

# Build and start all services
docker compose up --build -d

# Watch logs for errors (wait ~30s for services to stabilize)
docker compose logs --tail=50

# Verify all containers are running/healthy
docker compose ps

# Test the main service endpoint
curl -s http://localhost:${APP_PORT:-3000}/health || curl -s http://localhost:${APP_PORT:-3000}/

# If there's a frontend, check it too
curl -s http://localhost:${FRONTEND_PORT:-5173}/ | head -20

If something fails:

1. Check docker compose logs <service> for the failing service
2. Common issues: missing env vars, port conflicts, DB not ready before app starts
3. Fix and re-run docker compose up --build -d
4. Repeat until all services are healthy

Once verified, bring it down cleanly:

docker compose down

Step 6: Update AGENTS.md

Create or update AGENTS.md at the repository root with dev environment instructions. This file tells future agents how to spin up and use the environment.

If AGENTS.md doesn't exist, create it with this structure:

# AGENTS.md

## Dev Environment

### Prerequisites
- Docker and Docker Compose

### Starting the Environment
\`\`\`bash
docker compose up --build -d
\`\`\`

### Services
| Service | URL | Description |
|---------|-----|-------------|
| app     | http://localhost:3000 | Main application |
| db      | localhost:5432 | PostgreSQL database |

### Seeded Data
Describe what sample data is available (users, test accounts, etc.)

### Stopping the Environment
\`\`\`bash
docker compose down        # Stop containers
docker compose down -v     # Stop and remove volumes (reset DB)
\`\`\`

If AGENTS.md already exists, add a ## Dev Environment section (or update the existing one) with the same information. Do not remove existing content.

Also create a .env.example if one doesn't exist, documenting all environment variables used in docker-compose.yml with their defaults.

Troubleshooting