TCP vs HTTP health_check with terraform-google-lb

Question

I am using the terraform-google-lb (https://github.com/terraform-google-modules) module to create an http-https redirect with multiple backends. The module works great and all my backends and http redirect work. For whatever reason, the HTTP health_checks that the module creates will not work in my environment. When I change them manually to TCP health_checks, the health checks work. It looks like the module is hardcoded to create HTTP health_checks. Can this be changed to TCP health_checks with different TF code? The only way I can get this working so far is to delete the backend service (manually) and re-create it with a TCP health_check, which obviously subverts the automation considerably. The module uses this code (below) in main.tf , which apparently calls back into the module somewhwere and creates an HTTP hc. I would like to continue using the module because it streamlines many components that I would have to create piecemeal. But for the health_check, everything else works great.

locals {
  health_check = {
    check_interval_sec  = "5"
    timeout_sec         = "5"
    healthy_threshold   = "2"
    unhealthy_threshold = "2"
    request_path        = "/"
    port                = 80
    host                = null
    logging             = null
  }

inside the module the locals is referenced to the create the hc:

backends = {
    default = {
      description                     = null
      protocol                        = "HTTP"
      port                            = 80
      port_name                       = "http"
      timeout_sec                     = 10
      connection_draining_timeout_sec = null
      enable_cdn                      = false
      security_policy                 = null
      session_affinity                = null
      affinity_cookie_ttl_sec         = null
      custom_request_headers          = null
      custom_response_headers         = null
      health_check = local.health_check