AWS CloudFormation vs kubeconform

AWS CloudFormation models, provisions, and manages related AWS and third-party resources as infrastructure as code. YAML or JSON templates define AWS resources and their properties. StackSets deploy a common set of resources across multiple AWS accounts and Regions; Change Sets preview proposed stack updates, including possible deletion or replacement, and Drift Detection tracks changes made outside the service. For authoring, AWS CDK supports TypeScript, Python, Java, and .NET, while AWS SAM offers shorthand YAML for serverless resources that is converted during deployment. The CloudFormation Registry supports third-party resources and modules, and stacks can be updated from remote Git repositories. The service is available through API, Linux, macOS, web, and Windows, and manages state. CloudFormation itself has a free plan, with no additional charge for AWS and Alexa resource providers, but created AWS resources are billed at their standard rates. Third-party provider and custom hook operations may also incur charges. Access can be managed with IAM, and CloudTrail logs API calls.

Kubeconform is a command-line tool for validating Kubernetes manifests before deployment. It checks files against Kubernetes OpenAPI specifications, helping catch configuration problems before they reach a cluster. Validation and schema downloads run concurrently, and downloaded files are cached in memory. By default, it uses a self-updating fork of the Kubernetes JSON schema registry to supply schemas for recent Kubernetes versions. For custom resources, users can point it to local or remote schemas; local schemas also make offline validation possible. Results can be produced as JSON, JUnit, TAP, or plain text. Documentation describes using its Docker images in GitHub Actions and GitLab CI, and the images are published to GitHub Container Registry. It is free, with a listed plan of free. Installation instructions cover Linux, macOS through Homebrew, and Windows. Its checks do not cover additional server-side validation, and its Go module interface is still work in progress and unstable.