[Bug]: lexv2models/intent: `confirmation_setting` produces diffs for `prompt_attempts_specification`

[YakDriver]

Terraform Core Version

1.6.6

AWS Provider Version

5.32.1

Affected Resource(s)

• aws_lexv2models_intent
• aws_lexv2models_slot

Expected Behavior

No differences reported.

Actual Behavior

This is noted in the documentation.

Differences reported when you use confirmation_setting without configuring confirmation_setting.prompt_specification.prompt_attempts_specifications for Initial and Retry1.

Relevant Error/Panic Output Snippet

No response

Terraform Configuration Files

Reports diffs:

resource "aws_lexv2models_intent" "example" {
  bot_id      = aws_lexv2models_bot.test.id
  bot_version = aws_lexv2models_bot_locale.test.bot_version
  name        = "botens_namn"
  locale_id   = aws_lexv2models_bot_locale.test.locale_id

  confirmation_setting {
    active = true

    prompt_specification {
      allow_interrupt            = true
      max_retries                = 1
      message_selection_strategy = "Ordered"
    }
  }
}

Doesn't report diffs:

resource "aws_lexv2models_intent" "example" {
  bot_id      = aws_lexv2models_bot.test.id
  bot_version = aws_lexv2models_bot_locale.test.bot_version
  name        = "botens_namn"
  locale_id   = aws_lexv2models_bot_locale.test.locale_id

  confirmation_setting {
    active = true

    prompt_specification {
      allow_interrupt            = true
      max_retries                = 1
      message_selection_strategy = "Ordered"

      prompt_attempts_specification {
        allow_interrupt = true
        map_block_key   = "Initial"

        allowed_input_types {
          allow_audio_input = true
          allow_dtmf_input  = true
        }

        audio_and_dtmf_input_specification {
          start_timeout_ms = 4000

          audio_specification {
            end_timeout_ms = 640
            max_length_ms  = 15000
          }

          dtmf_specification {
            deletion_character = "*"
            end_character      = "#"
            end_timeout_ms     = 5000
            max_length         = 513
          }
        }

        text_input_specification {
          start_timeout_ms = 30000
        }
      }

      prompt_attempts_specification {
        allow_interrupt = true
        map_block_key   = "Retry1"

        allowed_input_types {
          allow_audio_input = true
          allow_dtmf_input  = true
        }

        audio_and_dtmf_input_specification {
          start_timeout_ms = 4000

          audio_specification {
            end_timeout_ms = 640
            max_length_ms  = 15000
          }

          dtmf_specification {
            deletion_character = "*"
            end_character      = "#"
            end_timeout_ms     = 5000
            max_length         = 513
          }
        }

        text_input_specification {
          start_timeout_ms = 30000
        }
      }
    }
  }
}

Steps to Reproduce

1. terraform apply

Debug Output

No response

Panic Output

No response

Important Factoids

No response

References

• 34891

Would you like to implement a fix?

None

[github-actions[bot]]

Community Note

Voting for Prioritization

• Please vote on this issue by adding a 👍 reaction to the original post to help the community and maintainers prioritize this request.
• Please see our prioritization guide for information on how we prioritize.
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.

Volunteering to Work on This Issue

• If you are interested in working on this issue, please leave a comment.
• If this would be your first contribution, please review the contribution guide.

[nam054]

Relates to #36845

[nam054]

This issue is a result of AWS returning more configurations than what was sent, which the provider cannot suppress the differences for at this time. Often these differences would be suppressed using custom diff suppress logic, but this behavior manifests in such a way that the upstream SDK (terraform-plugin-framework) does not allow this. Currently, the workaround is leaning on documentation (more detail here) that will exemplify exactly what is needed when dealing with the max_retries attribute. This is an open issue that details the same problem in framework: https://github.com/hashicorp/terraform-plugin-framework/issues/883.