data.kubernetes_ingress does not return status even when the Ingress is fully deployed

[brokedba]

Terraform Version, Provider Version and Kubernetes Version

Terraform version: v1.5.x
Kubernetes provider version: 2.22.0
Kubernetes version: 1.27.x

Affected Resource(s)

• data.kubernetes_ingress

Terraform Configuration Files

resource "helm_release" "open_webui" {
  name = "open-webui"

  repository = "https://helm.openwebui.com/"
  chart      = "open-webui"
  # version    = var.webui_chart_version # Optional: pin the chart version if needed
  wait       = true

  namespace        = "open-webui"
  create_namespace = true
  # Include the Helm values file
  values = [
    file("helm/webui_values.yaml") # Ensures relative path resolution
  ]
  # Set timeout to 15 minutes
  timeout = 900
}
###############
# check ingress
###############
 
# resource "time_sleep" "wait_for_ingress" {
#   depends_on = [helm_release.open_webui]
#   create_duration = "120s" # Adjust as necessary for your environment
# }

resource "null_resource" "wait_for_ingress" {
  triggers = {
    timestamp = timestamp()
  }

  depends_on = [helm_release.open_webui]
}
# Get the load balancer DNS
data "kubernetes_ingress" "openwebui_ingress" {
  metadata {
    name      = "open-webui"
    namespace = "open-webui"
  }
    depends_on = [null_resource.wait_for_ingress]
  
}

output "loadbalancer_dns" {
  value = {
    hostname = data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.hostname
    ip       = data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.ip
  }
}

resource "kubectl_manifest" "patched_openwebui_ingress" {
  yaml_body = <<EOT
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: open-webui
  namespace: open-webui
annotations:
  kubernetes.io/ingress.class: "traefik"
  cert-manager.io/cluster-issuer: "self-signed-cluster-issuer"
spec:
  rules:
    - host: "${data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.hostname}"
      http:
        paths:
          - path: /
            pathType: Prefix
            backend:
              service:
                name: open-webui
                port:
                  number: 8080
  tls:
    - hosts:
        - "${data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.hostname}"
      secretName: openwebui-tls-secret
EOT
  depends_on = [data.kubernetes_ingress.openwebui_ingress, kubectl_manifest.openwebui_certificate]
}`

Debug Output

Gist.--> ingress_datasource_deploy_failure.log

Panic Output

N/A

Steps to Reproduce

Expected Behavior

The data.kubernetes_ingress resource returns null for the status field, even though the Kubernetes Ingress resource is fully deployed and the kubectl command confirms the status.loadBalancer.ingress field is populated.

Actual Behavior

What actually happened?

helm_release.open_webui: Still creating... [2m30s elapsed]
helm_release.open_webui: Still creating... [2m40s elapsed]
helm_release.open_webui: Creation complete after 2m43s [id=open-webui]
null_resource.wait_for_ingress: Creating...
null_resource.wait_for_ingress: Creation complete after 0s [id=9110468478181892351]
data.kubernetes_ingress.openwebui_ingress: Reading...
data.kubernetes_ingress.openwebui_ingress: Read complete after 0s
╷
│ Error: Attempt to index null value
│
│   on helm_cert_manager.tf line 79, in resource "kubectl_manifest" "openwebui_certificate":
│   79:     - "${data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.hostname}"
│     ├────────────────
│     │ data.kubernetes_ingress.openwebui_ingress.status is null
│
│ This value is null, so it does not have any indices.
╵
╷
│ Error: Attempt to index null value
│
│   on helm_ollama_webui_notls.tf line 79, in output "loadbalancer_dns":
│   79:     hostname = data.kubernetes_ingress.openwebui_ingress.status.0.load_balancer.0.ingress.0.hostname
│     ├────────────────
│     │ data.kubernetes_ingress.openwebui_ingress.status is null
│
│ This value is null, so it does not have any indices.
╵

Important Factoids

The kubectl command verifies that the Ingress status.loadBalancer.ingress field is populated correctly:

kubectl get ingress open-webui -n open-webui -o yaml
Example Output:
status:
  loadBalancer:
    ingress:
    - hostname: valid-hostname.lb.civo.com
      ip: 192.0.2.1

Introducing delays (e.g., time_sleep) or dependencies on managed resources (e.g., helm_release) does not resolve the issue. The cluster is deployed in Civo Kubernetes, but the issue appears unrelated to the cluster provider.