create_before_destroy on aws_lambda_permission resources causing cycle errors

[mleziva]

Description

Occasionally seeing complex cycle errors during the plan phase that are difficult to debug. These errors are always related to the allowed_triggers input to the function as they can be resolved by removing that input. I have not been able to reliable reproduce this issue.

Versions

• Module version [Required]: 7.7.1

• Terraform version: 1.9.3

• Provider version(s): 5.64.0

Why is create_before_destroy being enforced on the aws_lambda_permission resources? If it is not required can we remove it or add other resources which do not have that lifecycle attribute?

[mleziva]

Based on my research, even if a resource is not created with a count or for_each attribute, if that resource has the lifecycle create_before_destroy = true, then all its dependencies will inherit the same.

My recommendation would be to not include an create_before_destroy resources in this module