Defang vs Direct Cloud Run Deployment

You have a multi-service Docker Compose application that works locally. You run docker compose up and your web server, API, database, and cache all come up together, discovering each other by name.

Cloud Run deploys individual containers, not Docker Compose applications. Each service becomes a separate Cloud Run service with its own configuration, networking, and deployment. Defang bridges this gap — it reads your compose.yaml and handles all the Cloud Run plumbing automatically.

The Docker Compose Gap

When you deploy a multi-service Docker Compose app to Cloud Run directly, you lose the things that make Docker Compose work:

Docker Compose Feature	Direct Cloud Run	Defang
Multi-service compose.yaml	Deploy each service separately	Deploys entire stack together
Service discovery by name	Not built in — use URLs	Automatic via private networking
depends_on ordering	Not supported	Respected
Single deploy command	One gcloud run deploy per service	defang compose up
Local-production parity	Different configs for each environment	Same compose.yaml everywhere
Environment variables	Set per-service in Console or CLI	Defined once in compose.yaml

Deployment Complexity

Direct Cloud Run

To deploy a typical Docker Compose web application with a database to Cloud Run, you need to:

1. Set up gcloud CLI and authenticate
2. Enable APIs: Cloud Run, Cloud Build, Artifact Registry, Cloud SQL, Secret Manager
3. Create Artifact Registry repository for your images
4. Configure Cloud Build to build your images
5. Create VPC and configure VPC Connector for Cloud SQL access
6. Set up Cloud SQL instance with proper networking
7. Create secrets in Secret Manager for database credentials
8. Deploy each Cloud Run service separately with environment variables, secrets, VPC connector
9. Configure IAM roles for service accounts

That’s 9+ manual steps per service, each with its own configuration options and potential for error.

With Defang

# Your entire Docker Compose stack
defang compose up --provider=gcp

Defang reads your compose.yaml and handles all the above automatically — for every service in your stack.

Feature Comparison

Feature	Defang	Direct Cloud Run
Docker Compose support	Native — deploys your compose.yaml	None — individual containers only
Multi-service deploy	All services deploy together	Manual per service
Database provisioning	One annotation (x-defang-postgres)	Manual Cloud SQL setup
Secret management	Automatic	Manual Secret Manager
VPC configuration	Automatic	Manual setup required
IAM configuration	Automatic	Manual roles needed
Local dev parity	Same compose.yaml locally and in production	Different configs per environment
LLM provisioning	One annotation (x-defang-llm)	Manual Vertex AI setup
MongoDB	One annotation (x-defang-mongodb)	Manual Firestore setup
Environments	Named stacks (dev/staging/prod)	Manual per-project

When to Use Each

Choose Defang when:

• You have a Docker Compose application and want to deploy it as-is
• You’re deploying multi-service applications
• You want managed database provisioning from your compose.yaml
• You value developer experience and local-production parity
• You want to minimize GCP-specific knowledge needed
• You want managed LLMs (Vertex AI) with a simple annotation
• You need multiple deployment environments (dev/staging/prod)

Choose Direct Cloud Run when:

• You need very specific Cloud Run configurations not available through Compose
• You’re integrating with complex existing GCP infrastructure
• You have dedicated DevOps resources and GCP expertise
• You’re deploying a single container, not a multi-service app
• You need features not yet supported by Defang

Try It

Take the compose.yaml you use with docker compose up locally and deploy it to GCP:

# Install Defang
brew install defang-io/defang/defang

# Deploy your existing compose.yaml to GCP
defang compose up --provider=gcp

Same Docker Compose file. Production Cloud Run infrastructure. No manual GCP configuration.