Terraform rollback

We have a situation where we moved the code from root module to child module which is causing the destruction/creation of some resources.

so,we would like to do the following -

  1. Create the new resource (endpoint in this case)first, test it for few days and ensure everything is working fine.

  2. If everything is working as expected, only then delete the old resource.

  3. If new resource is not working as expected or failed for some reason,roll-back the changes and continue to use the old resource.

But,our terraform code is trying to delete the old resource first and then creating new.I know we can use something like create_before_destroy argument but We would like to keep the old resource for few days to ensure things are
working as expected with new resources.

Any inputs would he appreciated.

The code is the description of what the current state should be, so I’d suggest changing the code in multiple steps.

Firstly create the new resource, then a few days later change the code to remove the old resource.