[proposal] Warn google_bigtable_instance without prevent_destroy

[wata727]

See https://www.terraform.io/docs/providers/google/r/bigtable_instance.html

Note: It is strongly recommended to set lifecycle { prevent_destroy = true } on instances in order to prevent accidental data loss. See Terraform docs for more information on lifecycle parameters.

[PatMyron]

Love the idea of encouraging explicit prevent_destroy values for stateful resource types: #25, #30, #37, hashicorp/terraform#24658, aws-cloudformation/cfn-lint#1232

Few thoughts:

1. One rule could cover the indefinitely expanding list of resource types, here's a similar expanding list for a similar rule:
   https://github.com/aws-cloudformation/cfn-lint/blob/main/src/cfnlint/data/AdditionalSpecs/StatefulResources.json
2. Useful for other providers like aws and azurerm as well
3. Don't think we should enforce a certain value of prevent_destroy itself, think we should just encourage explicitness alone

[PatMyron]

prevent_destroy also mentioned in https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/compute_disk

[wata727]

The reason I haven't worked on this is because I was worried that declaring a prevent_destroy for a test instance, could seem redundant. I want to avoid warnings for code that works correctly as much as possible.

However, I agree that it is good practice to always declare prevent_destroy so that it can be explicitly declared whether it is for testing that may be deleted or data that should not be deleted.