Genesys-cloud terraform provider crashes on full org export with large routing configurations

Trying to do a full org export for disaster recovery using the Genesys Cloud Terraform provider. The goal is to get a static snapshot of the entire configuration to store in a git repo. Running terraform init and terraform import works fine for individual resources, but the bulk export feature via the genesys_cloud provider’s data source or the CLI tool provided by the SDK keeps failing.

The error happens when the provider tries to fetch routing configurations. Specifically, it chokes on the routing_queue and routing_skill_group resources. The API seems to time out or return a partial payload that the SDK can’t parse into the expected Terraform state.

Here’s the config I’m using for the provider:

provider "genesyscloud" {
 environment = "mypurecloud.ie"
 client_id = var.genesys_client_id
 client_secret = var.genesys_client_secret
}

data "genesyscloud_export" "full_org" {
 filter {
 resource_type = "routing_queue"
 }
 filter {
 resource_type = "routing_skill_group"
 }
 output_path = "./export"
}

When I run terraform apply -target=data.genesyscloud_export.full_org, I get this error in the logs:

Error: Failed to export resource routing_queue
│ 
│ with data.genesyscloud_export.full_org,
│ on main.tf line 10, in data "genesyscloud_export" "full_org":
│ 10: data "genesyscloud_export" "full_org" {
│ 
│ Request returned status code 429. Rate limit exceeded. Retrying...
│ Failed after 5 retries.

I’ve tried increasing the retry count and adding delays, but the 429 errors persist. The org has over 500 queues, so the bulk fetch is hitting the API limits hard. Is there a way to paginate or throttle the export request within the provider config? Or should I be using a different approach, like querying the /api/v2/routing/queues endpoint directly with the SDK and then converting the JSON to HCL manually?

The SDK documentation doesn’t mention a bulk export method for the JavaScript client either. Just need a reliable way to get this done without hammering the API until it blocks us.