fix: add postgres service to docker-compose and wire backend dependency

[boys-cyberhub]

Closes #1

What changed

• Added postgres:15-alpine service to deployment/docker-compose.yml with credentials that match backend/.env.example defaults (user/password/stellar_portfolio)
• Added a pg_isready healthcheck on the postgres service so the backend waits until the DB is accepting connections before starting
• Added a redis-cli ping healthcheck to the existing redis service
• Updated backend depends_on to use condition: service_healthy for both postgres and redis — eliminates the race where the backend process launched before either dependency was ready
• Added DATABASE_URL to the backend environment pointing at postgres:5432 (the Docker service hostname) instead of localhost
• Added REDIS_URL to the backend environment (was already set in the monitoring service but missing from backend)
• Added postgres_data named volume so database state persists across restarts

Why

The backend connects to Postgres on module load and runs migrations at startup. Without a running Postgres service in the Compose stack the backend immediately exited with ECONNREFUSED 127.0.0.1:5432, making the entire Docker deployment path broken.

How to test

cd deployment
cp ../backend/.env.example .env
docker compose up
# All four services start cleanly; backend logs show migrations complete
curl http://localhost:3001/api/health   # 200 OK