### Terraform Version ```shell Terraform v1.5.7 ``` ### Terraform Configuration Files ```hcl terraform { backend "azurerm" {} } provider "google" { access_token = data.google_service_account_access_token.sa.access_token project = var.project_id region = "us-east1" } provider "google-beta" { access_token = data.google_service_account_access_token.sa.access_token project = var.project_id region = "us-east1" } provider "google" { credentials = file(var.gcp_credential_path) alias = "impersonated" } data "google_service_account_access_token" "sa" { provider = google.impersonated target_service_account = "sa-${substr(var.project_id, 0, min(length(var.project_id), 27))}@${var.project_id}.iam.gserviceaccount.com" lifetime = "3600s" scopes = [ "https://www.googleapis.com/auth/cloud-platform", ] } module "gcp_pubsub" { source = "acnciotfregistry.accenture.com/accenture-cio/pubsub/google" version = "2.1.0" project_id = var.project_id topic_name = "NewTopic-us-east1" subscription_name = "NewSubscription-us-east1" intended_event_type = "Application" resources_to_create = "BOTH" delivery_type = "Push" endpoint_url = "https://www.testendpointurl.com" push_auth_service_account = "sa-103879-big-data@${var.project_id}.iam.gserviceaccount.com" labels = { examplenum = "gcp_pubsub_test" creates = "subscription" } } ``` ### Debug Output NA ### Expected Behavior NA ### Actual Behavior Problem: The latest version of the Terraform Google Cloud provider has introduced a change in the default retry policy for Pub/Sub subscriptions. Previously, the default was "retry immediately," but now it defaults to "retry with exponential backoff" with a default backoff configuration (retry_policy_time = {minimum_backoff = "10s" maximum_backoff = "600s"}). This unexpected behavior is causing issues for our application teams, who require the "retry immediately" policy. When we hardcode version 6.0 of the Google Cloud provider in our latest terraform module, it automatically applied the "retry immediately" policy (no backoff was provided), it has worked as expected. Currently we have provided this workaround for most of our application teams. But we don't recommend to hardcode Google Cloud provider version in our modules as per our process standards. Impact: The default exponential backoff policy is causing delays in message delivery, impacting the performance and responsiveness of our applications. Requirement: We need to ensure that the default retry policy for Pub/Sub subscriptions reverts to "retry immediately" when using the latest version of the Terraform Google Cloud provider. Our Application teams requirement is "retry immediately" Please resolve this bug immediately ### Steps to Reproduce NA