Support for node_network_profile

[RelaxingLoki475]

Is there an existing issue for this?

• I have searched the existing issues

Description

Have the ability to manage node_network_profile to set application_security_group_ids in order to improve NSG management.

New or Affected Resource(s)/Data Source(s)

azurerm_kubernetes_cluster

Potential Terraform Configuration

No response

References

https://registry.terraform.io/providers/hashicorp/azurerm/latest/docs/resources/kubernetes_cluster#node_network_profile

[zioproto]

Documentation reference:
https://learn.microsoft.com/en-us/azure/aks/use-node-public-ips#use-public-ip-tags-on-node-public-ips-preview

[jramacpr]

I face the same issue when trying to attach an ASG with the node pool of the aks. Though terraform provides option to include the ASG inside the node network profile, it seems it do not work as expected.

  default_node_pool {
      name                   = var.agents_pool_name
      node_count             = var.agents_count
      vm_size                = var.agents_size
      vnet_subnet_id         = data.azurerm_subnet.subnet.id
      auto_scaling_enabled   = var.enable_auto_scaling
      max_count              = var.max_count
      min_count              = var.min_count
      node_public_ip_enabled = var.enable_node_public_ip
      max_pods               = var.agents_max_pods
      node_network_profile {
          application_security_group_ids = [data.azurerm_application_security_group.vm_asg.id]
      }
      upgrade_settings {
        max_surge = "10%"
      }
  }

After applying this ,the default node pool's vm scale set has no ASG's associated with it .Any manual association of the ASG is getting reverted to the original state(state from terraform config ,which sees as NO ASG's to be attached to the vm scale set).So as a result currently attaching the ASG with the node network profile is not working as expected

[zioproto]

@jramacpr I understand you are trying to pass the id of an application security group into the
default_node_pool block.

Could you confirm if you are using the module Azure/terraform-azurerm-aks or if you are using the azurerm_kubernetes_cluster resource directly ?

In the module it seems this feature is implemented only for the additional node pools but not for the default node pool:

terraform-azurerm-aks/extra_node_pool.tf

Line 117 in 495d0ab

dynamic "node_network_profile" {

[jramacpr]

@zioproto iam using azurerm_kubernetes_cluster and as per the document it says
A node_network_profile block supports the following:

application_security_group_ids - (Optional) A list of Application Security Group IDs which should be associated with this Node Pool.

[zioproto]

@jramacpr I opened PR #598 because it seems the module did not support node_network_profile in the default pool.

However, if you are consuming the resource azurerm_kubernetes_cluster directly you should open an issue at https://github.com/hashicorp/terraform-provider-azurerm/issues because this could be an issue with the upstream provider.

What version of the Hashicorp azurerm provider are you using in your project ?

thanks

[fuog]

I face the same issue when trying to attach an ASG with the node pool of the aks. Though terraform provides option to include the ASG inside the node network profile, it seems it do not work as expected.

  default_node_pool {
      name                   = var.agents_pool_name
      node_count             = var.agents_count
      vm_size                = var.agents_size
      vnet_subnet_id         = data.azurerm_subnet.subnet.id
      auto_scaling_enabled   = var.enable_auto_scaling
      max_count              = var.max_count
      min_count              = var.min_count
      node_public_ip_enabled = var.enable_node_public_ip
      max_pods               = var.agents_max_pods
      node_network_profile {
          application_security_group_ids = [data.azurerm_application_security_group.vm_asg.id]
      }
      upgrade_settings {
        max_surge = "10%"
      }
  }

After applying this ,the default node pool's vm scale set has no ASG's associated with it .Any manual association of the ASG is getting reverted to the original state(state from terraform config ,which sees as NO ASG's to be attached to the vm scale set).So as a result currently attaching the ASG with the node network profile is not working as expected

We have the same problem as you describe. Did you find any solution? is the provider the problem?