enh: allow multiple instance types for aws

This was squashed with the commit for

    fix: change check for multiple instance types

and while rebasing we had a conflict for
	src/_nebari/stages/infrastructure/template/aws/modules/kubernetes/main.tf

Annotation for our "manual patch queue":

PR: nebari-dev#3070

Author: satra

src/_nebari/stages/infrastructure/__init__.py

@@ -619,10 +619,11 @@ def _check_input(cls, data: Any) -> Any:
                     if hasattr(node_group, "__getitem__")
                     else node_group.instance
                 )
-                if instance not in available_instances:
-                    raise ValueError(
-                        f"Amazon Web Services instance {node_group.instance} not one of available instance types={available_instances}"
-                    )
+                for instance_name in instance.split(","):
+                    if instance_name not in available_instances:
+                        raise ValueError(
+                            f"Amazon Web Services instance {instance_name} not one of available instance types={available_instances}"
+                        )
 
         # check if kms key is valid
         available_kms_keys = amazon_web_services.kms_key_arns(data["region"])

src/_nebari/stages/infrastructure/template/aws/modules/kubernetes/main.tf

@@ -91,7 +91,7 @@ resource "aws_eks_node_group" "main" {
   node_role_arn   = aws_iam_role.node-group.arn
   subnet_ids      = var.node_groups[count.index].single_subnet ? [element(var.cluster_subnets, 0)] : var.cluster_subnets
 
-  instance_types = [var.node_groups[count.index].instance_type]
+  instance_types = split(",", var.node_groups[count.index].instance_type)
   capacity_type  = var.node_groups[count.index].spot ? "SPOT" : "ON_DEMAND"
   ami_type       = var.node_groups[count.index].ami_type
   disk_size      = var.node_groups[count.index].launch_template == null ? 50 : null