Deploy failure 와 오래된 코드로의 deploy

Based on your description, you're experiencing two main issues: deployment failures and the system reverting to an older version of your code.

For the deployment failures:

1. The error you're seeing could be related to a validation issue in your AWS CloudFormation template or custom resource. When deploying infrastructure as code, validation errors can cause deployments to fail and roll back to previous versions.

2. Check your CloudFormation stack status in the AWS Console. If it's in a ROLLBACK_FAILED state, you may need to manually delete the stack from the AWS console before attempting to deploy again.

3. If you're using AWS CDK or a similar tool, you might need to add output path specifications to limit the response size from custom resources, as large response objects can cause deployment failures.

For the reversion to older code:

1. Elastic Beanstalk might be rolling back to the last known good deployment when it encounters errors with your newer code.

2. Check your deployment logs to understand why the newer version is failing to deploy properly.

Regarding billing concerns:

1. Even when applications are in a "Severe" state, AWS resources remain provisioned and continue to incur charges.

2. Consider temporarily shutting down unnecessary resources while troubleshooting if cost is a concern.

To resolve these issues:

1. Review your deployment logs for specific error messages.

2. Validate your configuration files and application code.

3. If using custom resources in CloudFormation, ensure they're properly configured with appropriate output paths.

4. Consider using infrastructure as code tools like Terraform or AWS CDK for more predictable deployments.

5. If the issue persists, you may need to create a new environment and deploy your application there instead of trying to update the problematic one.
   Sources
   Deploying Stack throws error at CreateKeysAndCertificate | AWS re:Post
   𝗪𝗵𝘆 𝗧𝗲𝗿𝗿𝗮𝗳𝗼𝗿𝗺 𝗶𝘀 𝗮 𝗚𝗮𝗺𝗲-𝗖𝗵𝗮𝗻𝗴𝗲𝗿 | AWS re:Post