Add example of how to retrieve kubeconfig (#19)

managed-k8s/README.md

@@ -18,11 +18,38 @@ whereas each worker node has
 
 Provisioning time varies of course but we can expect to take around 10 minutes to spin up the cluster.
 
+## Retrieve kubeconfig
+
+Inside the cluster.tf you can see a defined output. the configured `value` needs
+to match the `gridscale_k8s` resource. In the given example, our `gridscale_k8s`
+resource is named `mycluster`:
+
+```hcl
+resource "gridscale_k8s" "mycluster" {
+  # ...
+}
+```
+
+So the configured value of our terraform `output` needs to look like this:
+
+```hcl
+output "kubeconfig" {
+  value = gridscale_k8s.mycluster.kubeconfig
+}
+```
+
+When you create your cluster terraform will write the kubeconfig now to stdout,
+if you need to get the kubeconfig _after_ you've deployed your cluster, you can
+do so by calling:
+
+```bash
+terraform output kubeconfig
+```
+
 ## TODO
 
 Missing:
 
-- Show how to retrieve kubeconfig
 - Run through a minimum hello world example:
   - create a deployment
   - create a service

managed-k8s/cluster.tf

@@ -13,7 +13,7 @@ resource "gridscale_k8s" "mycluster" {
   # Define the desired Kubernetes release. Define which release of GSK will be
   # used for the cluster. For convenience, please use gscloud to get the list of
   # available Kubernetes releases.
-  release = "1.17"
+  release = "1.21"
 
   node_pool {
     name         = "my_node_pool"
@@ -24,3 +24,7 @@ resource "gridscale_k8s" "mycluster" {
     storage_type = "storage_insane"
   }
 }
+
+output "kubeconfig" {
+  value = gridscale_k8s.mycluster.kubeconfig
+}