aws: change default controller model to t3.medium

As 2GB of RAM is not enough for running standalone controller node with
graceful upgrade strategy (2 API servers does not fit into 2GB node).

Signed-off-by: Mateusz Gozdek <mateusz@kinvolk.io>

assets/lokomotive-kubernetes/aws/flatcar-linux/kubernetes/variables.tf

@@ -30,8 +30,14 @@ variable "worker_count" {
 }
 
 variable "controller_type" {
-  type        = string
-  default     = "t3.small"
+  type = string
+  # When doing the upgrades of controlplane on t3.small instance type when
+  # having one single controlplane node, t3.small has not enough memory (2GB)
+  # to run more than one instance of kube-apiserver in parallel, so we need to use
+  # a bigger instance. With HA controlplane, t3.small should be fine, though for
+  # production setups, it's recommended to use bigger instances (like t3.medium), to
+  # give plenty of usable memory for etcd and kube-apiserver.
+  default     = "t3.medium"
   description = "EC2 instance type for controllers"
 }