Terraform - AKS Private Cloud | Infinite wait on helm relase - terraform

Question

I am trying to create a Private Cloud on AKS with Terraform.
The public route seemd to work fine and I am putting in security stuff, step by step
After putting in Networks azurerm_virtual_network, azurerm_subnet it seems to hand my Helm Deployment
There are no logs, its just an infinite wait.
helm_release.ingress: Still creating... [11m0s elapsed] (this is a simple NGINX Ingress Controller)
resource "azurerm_virtual_network" "foo_network" {
name = "${var.prefix}-network"
location = azurerm_resource_group.foo_group.location
resource_group_name = azurerm_resource_group.foo_group.name
address_space = ["10.1.0.0/16"]
}
resource "azurerm_subnet" "internal" {
name = "internal"
virtual_network_name = azurerm_virtual_network.foo_network.name
resource_group_name = azurerm_resource_group.foo_group.name
address_prefixes = ["10.1.0.0/22"]
}
Any points on how should I debug this? Lack of logs is making it difficult to understand.
Complete Script
provider "azurerm" {
features {}
}
resource "azurerm_resource_group" "foo" {
name = "${var.prefix}-k8s-resources"
location = var.location
}
resource "azurerm_kubernetes_cluster" "foo" {
name = "${var.prefix}-k8s"
location = azurerm_resource_group.foo.location
resource_group_name = azurerm_resource_group.foo.name
dns_prefix = "${var.prefix}-k8s"
default_node_pool {
name = "system"
node_count = 1
vm_size = "Standard_D4s_v3"
}
identity {
type = "SystemAssigned"
}
addon_profile {
aci_connector_linux {
enabled = false
}
azure_policy {
enabled = false
}
http_application_routing {
enabled = false
}
kube_dashboard {
enabled = true
}
oms_agent {
enabled = false
}
}
}
provider "kubernetes" {
version = "~> 1.11.3"
load_config_file = false
host = azurerm_kubernetes_cluster.foo.kube_config.0.host
username = azurerm_kubernetes_cluster.foo.kube_config.0.username
password = azurerm_kubernetes_cluster.foo.kube_config.0.password
cluster_ca_certificate = base64decode(azurerm_kubernetes_cluster.foo.kube_config.0.cluster_ca_certificate)
}
provider "helm" {
# Use provider with Helm 3.x support
version = "~> 1.2.2"
}
resource "null_resource" "configure_kubectl" {
provisioner "local-exec" {
command = "az aks get-credentials --resource-group ${azurerm_resource_group.foo.name} --name ${azurerm_kubernetes_cluster.foo.name} --overwrite-existing"
environment = {
KUBECONFIG = ""
}
}
depends_on = [azurerm_kubernetes_cluster.foo]
}
resource "helm_release" "ingress" {
name = "ingress-foo"
repository = "https://kubernetes.github.io/ingress-nginx"
chart = "ingress-nginx"
timeout = 3000
depends_on = [null_resource.configure_kubectl]
}

Answer 1

The best way to debug this is to be able to kubectl into the AKS cluster. (AKS should have documentation on how to set up kubectl.)
Then, play around with kubectl get pods -A and see if anything jumps out as being wrong. Specifically, look for nginx-ingress pods that are not in a Running status.
If there are such pods, debug further with kubectl describe pod <pod_name> or kubectl logs -f <pod_name>, depending on whether the issue happens after the container has successfully started up or not.