Support regional endpoints for google_ml_engine_model

[WouterFlorijn]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment. If the issue is assigned to the "modular-magician" user, it is either in the process of being autogenerated, or is planned to be autogenerated soon. If the issue is assigned to a user, that user is claiming responsibility for the issue. If the issue is assigned to "hashibot", a community member has claimed the issue already.

Description

Currently, the google_ml_engine_model resource does not support regional endpoints (a checkbox in the GCP console). It would be great if this was added through a boolean like regional_endpoint.

New or Affected Resource(s)

• google_ml_engine_model

Potential Terraform Configuration

resource "google_ml_engine_model" "my_model" {
  name = "my_model"
  description = "My model"
  regional_endpoint = true # Defaults to false.
  regions = ["europe-west4"]
}

Maybe region should be used instead of regions when regional_endpoint is set to true.

[AndreasBergmeier6176]

There is no real need. You can just do:

provider "google-regional" {
  alias = "regional_ml"
  project                   = "foo-bar"
  region                    = "europe-west2"
  ml_engine_custom_endpoint = "https://europe-west2-ml.googleapis.com/v1/"
}

[AarshDhokai]

b/262230505

[AarshDhokai]

I've looked into this issue and the following are some of the details that I've found out:

• In this google_ml_engine_model resource, regional endpoints are supported.
• But to provide support for this feature in the google_ml_engine_model resource, we need to provide support for 2 different base paths (https://ml.googleapis.com/ and https://{{region}}-ml.googleapis.com ) and that seems to be too rigid to handle with MMv1.
• The probable solution to this could be to write a new regional handwritten resource that can support a base path https://{{region}}-ml.googleapis.com. Alternatively we can try this approach which uses alias and ml_engine_custom_endpoint attribute for different providers and uses it for regional endpoints for google_ml_engine_model resource.