Instance groups from multi zone cluster about terraform-google-lb-http HOT 4 CLOSED

I am not familiar with the issue that you're describing but I have not worked extensively with this module. If you think there is work needed on the module after hashicorp/terraform-provider-google#3937 is addressed then please do open a new issue here.

I'm closing this issue since you've managed to answer your own question. 😄

from terraform-google-lb-http.

Any info on this one?

from terraform-google-lb-http.

Hi @hawksight and @mkozjak. Thank you for your interest in the project! I apologize for the extreme delay in responding.

You can specify multiple backend objects for each index in the backends input, but each backend object can only specify one group. I think it should be possible to dynamically build the backends input using for expressions.

from terraform-google-lb-http.

Thanks @aaron-lane - I've been using this module as the basis of my own for a long time.
Being able to loop through the backends so that the number of them is not longer hard coded is now possible for creation using a dynamic block.

Also to answer my own original question - yes this module will support multi zone clusters and cluster with multiple node pool. You just have to ensure you get the complete list of instance_groups in order to populate the group value of the backends input of this module.

The same limitation seems to apply that you will need to know up front how many instance_groups your going to have. In reality I find I do know as we are fairly static in the group respect. Your mileage may vary.

I just checked through this module's recent code for your 0.12 TF support where you use a dynamic block here similar to what I have in my own module

  dynamic "backend" {
    for_each = var.backend_group_list
    content {
      group = backend.value
      // adding null values otherwise reapplication fails
      balancing_mode               = null
      capacity_scaler              = null
      description                  = null
      max_connections              = null
      max_connections_per_instance = null
      max_rate                     = null
      max_rate_per_instance        = null
      max_utilization              = null
    }
  }

Where backend_group_list = data.google_container_cluster.information.instance_group_urls.
I set null here so that the default apply without having to know them.

Have you seen any problems updating load balancers after they are initially built? I find my code currently works to build, but tf apply on a plan once the backend_service resources are built comes back with errors. I've got an issue open here but wondered if you might have come across the errors I've seen when updating the module?

from terraform-google-lb-http.