feat: support Data API for provisioned Redshift clusters

[alexjurkiewicz]

Summary

The data_api provider block previously only supported Redshift Serverless workgroups. This PR extends it to also support provisioned clusters via a cluster_identifier + username pair.

The underlying redshift-data-sql-driver already supports both DSN formats — this change wires up the provider schema and config-building logic to expose that capability.

Changes

• New cluster_identifier field in data_api block (mutually exclusive with workgroup_name via ExactlyOneOf)
• New username field in data_api block (the DB user passed as DbUser in ExecuteStatementInput; required at apply time when cluster_identifier is set)
• workgroup_name changed from Required to Optional (schema enforces exactly one of workgroup_name/cluster_identifier must be set)
• temporary_credentials added to data_api's ConflictsWith to make the conflict symmetric
• New example: examples/provider/provider_using_redshift_data_api_provisioned.tf

Example usage

provider "redshift" {
  database = "mydb"
  data_api {
    cluster_identifier = "my-cluster"
    username           = "myuser"
    region             = "us-east-1"
  }
}

Testing

• Unit tests added for DSN building, happy path, missing-username error, and ResourceData-level schema→config wiring
• Manually tested end-to-end against a live provisioned cluster via the Data API (create and destroy of a redshift_user)
• Existing workgroup path unchanged (regression test included)

🤖 Generated with Claude Code

[alexjurkiewicz]

On a personal note -- thank you for keeping the "Redshift user management with Terraform" dream alive 😊