Terraform vs CloudFormation

Terraform and AWS CloudFormation are the two most common IaC tools for AWS infrastructure. Terraform is multi-cloud and uses HCL. CloudFormation is AWS-native, uses JSON/YAML, and has deep AWS integration with zero state management overhead.

Multi-Cloud vs AWS-Native

The biggest differentiator is scope. Terraform can manage resources across AWS, GCP, Azure, Kubernetes, Datadog, PagerDuty, and thousands of other providers. CloudFormation is exclusively for AWS.

If your infrastructure is entirely on AWS and will remain so, CloudFormation's native integration is compelling. If you have multi-cloud needs or want a single IaC tool for everything, Terraform is the clear choice.

State Management

CloudFormation manages state on the AWS side automatically. You never deal with a state file, locking, or remote backends. This eliminates an entire class of operational problems.

Terraform requires you to configure and manage a remote state backend (typically S3 + DynamoDB for locking). State file corruption or conflicts can be painful, though Terraform Cloud simplifies this.

Rollback Behavior

CloudFormation automatically rolls back failed stack operations, returning resources to their previous state. This safety net is valuable for production deployments.

Terraform has no automatic rollback. If a terraform apply fails midway, you may end up in a partially applied state that requires manual intervention.

Developer Experience

HCL is significantly more concise and readable than CloudFormation's JSON/YAML. A resource that takes 5 lines in HCL might take 20 lines in CloudFormation YAML. Terraform modules are more ergonomic than CloudFormation's nested stacks.

For AWS architecture patterns, explore our concepts and system design interview guide. Practice with interview questions and see pricing for premium access.