How do lifecycle rules change create, update and destroy behaviour?

Lifecycle settings apply to one resource block. They do not change the provider API, only Terraform's ordering and safety checks. Used well, they prevent outages and accidental deletion. Used broadly, they hide real drift and create replacement conflicts.

resource "aws_instance" "app" {
  ami           = data.aws_ami.base.id
  instance_type = "t3.micro"

  lifecycle {
    create_before_destroy = true
    prevent_destroy       = true
    ignore_changes        = [user_data]
  }
}

The main rules

• create_before_destroy reverses the default destroy-then-create order
• prevent_destroy fails the plan when a destroy would occur
• ignore_changes keeps specified attributes from causing updates
• replace_triggered_by forces replacement when another resource changes
• precondition and postcondition validate assumptions during apply

create_before_destroy can fail when names must be unique, because the old and new objects coexist briefly. It also interacts with dependencies that still point at the old object. ignore_changes can mask drift permanently, so scope it to computed or externally managed attributes and revisit it.