Terraform aws downtime during apply

Question

I create some resources with terraform: EC2, target group, autoscaling roup, load balancer and more... When I change some paramas, for example instance type, the current target group drain the targeted EC2 before the new EC2 are added. So I have a downtime. How can I change my terraform conf to prevent downtime?

In all resource I use:

  lifecycle {
    create_before_destroy = true
  }

This is the module I call to create all resource I needed:

module "autoscaling_docker" {
    source = "../../_/autoscaling_docker"

    domain        = local.domain
    has_keychain  = true
    instance_type = "t3.medium"
    name          = "cluster-name"
... 

The module autoscaling_docker create various resources like autoscaling group (and so ec2), alb and many more

This is my terraform target group:

resource "aws_lb_target_group" "me" {
  name        = "tg-tf-${var.name}"
  port        = 80
  protocol    = "HTTP"
  vpc_id      = var.vpc_id
  target_type = "instance"

  stickiness {
    type            = "lb_cookie"
    cookie_duration = "300"
  }

  health_check {
    enabled             = true
    healthy_threshold   = 2
    interval            = 30
    matcher             = "200"
    path                = var.health_check_path
    port                = 80
    protocol            = "HTTP"
    timeout             = 5
    unhealthy_threshold = 2
  }

  lifecycle {
    create_before_destroy = true
  }

  deregistration_delay = "360"
}

Autoscaling group:

resource "aws_autoscaling_group" "me" {
...
lifecycle {
    ignore_changes = [
      desired_capacity
    ]
    create_before_destroy = true
  }