Best practices

Use terraform_data for custom scripts and provisioners.

Use triggers to control resource recreation.

Implement a local-exec provisioner for complex setup tasks.

Use depends_on for explicit dependency management.

Store state for idempotent operations.

Implement proper error handling in scripts.

Validate and sanitize inputs for security.

Log output for troubleshooting.

Use connection blocks for remote execution.

When you implement features with terraform_data, consider it a temporary solution. Regularly check if the provider has added native support for that functionality. When the provider introduces native capability for your terraform_data implementation, switch to the provider's official solution.

Don't overuse terraform_data as a primary resource management tool.

Avoid storing sensitive information directly in scripts.

Don't ignore error handling in provisioners.

Avoid complex, long-running scripts. Don't create null resources with multiple responsibilities and excessively complex workflows.

Don't create unnecessary dependencies. Follow state management principles and add only necessary triggers and depends_on conditions in your code.

Don't use terraform_data for core infrastructure management. Don't replace your standard resource management with custom resources.

Don't hardcode credentials or sensitive data.

Don't ignore security best practices.

Don't skip input validation.