Add rolling_upgrade_policy to azurerm_linux_virtual_machine_scale_set

[0x46616c6b]

What problem would you like to address?

I would like to add the ability to configure rolling_upgrade_policy in azurerm_linux_virtual_machine_scale_set. This will solve some unavailabilities of TFE when OS updates are applied.

Should I create a PR for that change?

Who is the audience for this request?

TFE Customer

What would the ideal solution look like?

Have a configurable rolling_update_policy with some defaults:

resource "azurerm_linux_virtual_machine_scale_set" "tfe_vmss" {
  // Other options in the resource

  dynamic "rolling_upgrade_policy" {
    for_each = var.vm_upgrade_mode == "Rolling" ? [1] : []

    content {
      cross_zone_upgrades_enabled             = false
      max_batch_instance_percent              = 20
      max_unhealthy_instance_percent          = 20
      max_unhealthy_upgraded_instance_percent = 20
      pause_time_between_batches              = "PT0S"
      prioritize_unhealthy_instances_enabled  = false
    }
  }
}

var.vm_upgrade_mode defaults to "Manual", so the block wouldn't apply.

What alternatives have you considered?

Nothing.

How does this impact you?

I inlined the vm module for my TFE installation and applied this change already.

Are there any deadlines?

No.

Is there any additional context?

No.