Could someone explain why genesyscloud_wfm_user_schedule resources are failing with 422 Unprocessable Entity during a large-scale apply in our CI/CD pipeline? We are managing schedule imports for ~500 agents using a for_each loop driven by CSV data. The pipeline uses Terraform v1.9.8 and the Genesys Cloud provider v1.65.0. The error response body contains {"code": "conflict", "message": "Schedule segment overlaps with existing segment", "errors": [{"message": "Cannot add schedule segment due to overlap"}]}. This is confusing because we run terraform plan with -refresh=false to ensure state matches the remote API before applying, and the plan shows no changes to existing segments. We are also using depends_on to ensure the genesyscloud_wfm_user and genesyscloud_wfm_schedule_group exist before applying schedules. The issue appears intermittent, affecting only 10-15% of the apply run. We suspect rate limiting or eventual consistency issues with the WFM API endpoints (/api/v2/wfm/users/{userId}/schedule). We have tried adding time_sleep resources between batches of 50 users, but the 422 errors persist. The JSON payload generated by Terraform matches the OpenAPI spec for POST /api/v2/wfm/users/{userId}/schedule. We are based in Australia/Sydney, so timezone handling is critical. The start and end timestamps in our HCL are UTC formatted. Is there a known issue with the provider’s retry logic for WFM schedule conflicts? Or should we implement a custom retry mechanism in our GitHub Actions workflow using genesyscloud_wfm_user_schedule data sources to validate state before apply? We need a reliable way to automate schedule deployments without manual intervention in the Admin UI.