examples/kubernetes can't be modified to use a static cluster name

[kriswuollett]

To avoid the data lookup, should the example be modified to use the dependency inversion as described in the Terraform docs? For example, just have the main module pass a basic primary_cluster value between the submodules and remove the data "digitalocean_kubernetes_cluster" "primary" from the kubernetes-config submodule:

examples/kubernetes/doks-cluster/outputs.tf:

output "primary_cluster" {
  value = {
    id                     = digitalocean_kubernetes_cluster.primary.id
    name                   = digitalocean_kubernetes_cluster.primary.name
    endpoint               = digitalocean_kubernetes_cluster.primary.endpoint
    token                  = digitalocean_kubernetes_cluster.primary.kube_config[0].token
    cluster_ca_certificate = digitalocean_kubernetes_cluster.primary.kube_config[0].cluster_ca_certificate
    raw_config             = digitalocean_kubernetes_cluster.primary.kube_config[0].raw_config
  }
}

examples/kubernetes/kubernetes-config/variables.tf:

variable "primary_cluster" {
  type = object({
    id                     = string
    name                   = string
    endpoint               = string
    token                  = string
    cluster_ca_certificate = string
    raw_config             = string
  })
}

Discussed in #690

^{Originally posted by kriswuollett September 13, 2021}
Is there a particular reason why cluster_name being tf-k8s-${random_id.cluster_name.hex} was needed in #564? When I remove the random_id resource and just set cluster_name to my_cluster, the output of terraform plan has the error:

│ Error: Unable to find cluster with name: my-cluster
│ 
│   with module.kubernetes-config.data.digitalocean_kubernetes_cluster.primary,
│   on kubernetes-config/main.tf line 18, in data "digitalocean_kubernetes_cluster" "primary":
│   18: data "digitalocean_kubernetes_cluster" "primary" {

I'm new to Terraform so would like to learn more about this issue if it is not technically a bug. Thanks!