Describe the bug
tailscale_device_subnet_routes resource cannot be updated after being previously applied to a now deleted ephemeral device.

To Reproduce
Steps to reproduce the behaviour:

• Apply Terraform code.

resource "tailscale_tailnet_key" "this" {
  reusable    = true
  expiry = 90 * 24 * 60 * 60 # 90 days
  ephemeral   = true
  description = "Kubernetes Cluster"
  tags = ["tag:k8s"]
}

# ... k8s pod definition ...

data "tailscale_devices" "ts_devices" {
  name_prefix = "tailscale-subnet-router-"

  depends_on = [kubernetes_deployment_v1.this]
}

resource "tailscale_device_subnet_routes" "routes" {
  device_id = data.tailscale_devices.ts_devices.devices[0].node_id
  routes = [
    "10.42.0.0/16",
    "10.43.0.0/16",
    "192.168.1.0/24",
    # Configure as an exit node
    "0.0.0.0/0",
    "::/0"
  ]
}
...

• Delete the Pod in the Kubernetes cluster. A new pod is created with a new hostname.

• Attempt to apply the Terraform again.

• Received the following error:

Error: Failed to fetch device subnet routes
no manageable device matching this ID found (404)

Expected behaviour
tailscale_device_subnet_routes resource should be marked for replacement when device_id field changes.

Additionally, fail gracefully if the device doesn't exist in Tailscale anymore.

Desktop (please complete the following information):

• OS: Mac OS X 15.5
• OpenTofu Version 1.9.0
• Provider Version 0.21.1