Why does this setting trigger a 400 error when deploying via Terraform on AU-1 BYOC?
Environment details are consistent with previous deployments. Provider version is v1.22.0. The issue arises specifically with the genesyscloud_ai_conversation_config resource. We are attempting to automate the deployment of AI bot configurations using Terraform. The goal is to set the intent_confidence_threshold dynamically based on environment variables.
The HCL configuration looks standard. No syntax errors. The plan phase completes successfully. The apply phase fails immediately on the resource creation.
resource "genesyscloud_ai_conversation_config" "main_bot" {
name = "Production Bot Config"
description = "Managed by Terraform"
intent_confidence_threshold = var.bot_confidence_threshold
language_pack_id = "en-US"
}
The error response from the API is not descriptive. It returns a generic 400 Bad Request. No specific field validation error is included in the JSON response body. This makes debugging difficult.
We suspect it might be related to the language_pack_id or the threshold format. The variable var.bot_confidence_threshold is passed as a string “0.75”. We have tried passing it as a number 0.75. Both result in the same 400 error.
Other AI resources like genesyscloud_ai_knowledge_base deploy without issues. This suggests the problem is isolated to the conversation config endpoint or the Terraform provider implementation for this specific resource.
Has anyone successfully deployed genesyscloud_ai_conversation_config with dynamic threshold values? Is there a known limitation with the AU-1 region for this resource? We need to move forward with the CI/CD pipeline automation. Manual configuration is not an option for our scale. Any insights into the expected payload format for this API endpoint would be appreciated. The documentation is sparse on the specific validation rules for the confidence threshold field.