google_vertex_ai_index documents requied fields as optional

[natedogith1]

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 an issue is assigned to a user, that user is claiming responsibility for the issue.
• Customers working with a Google Technical Account Manager or Customer Engineer can ask them to reach out internally to expedite investigation and resolution of this issue.

Terraform Version

Terraform v1.3.9 on linux_amd64

Affected Resource(s)

google_vertex_ai_index

Terraform Configuration

resource "google_vertex_ai_index" "index" {
  project             = var.project
  display_name        = local.display_name
  description         = var.description
  index_update_method = var.index_update_method
  region              = var.region
  labels              = var.labels
  metadata {
    contents_delta_uri    = var.contents_delta_uri
    is_complete_overwrite = var.is_complete_overwrite 
    config {
      dimensions                  = var.metadata_config.dimensions
      approximate_neighbors_count = var.metadata_config.approximate_neighbors_count
      shard_size                  = var.metadata_config.shard_size
      distance_measure_type       = var.metadata_config.distance_measure_type
      feature_norm_type           = var.metadata_config.feature_norm_type
      algorithm_config {}
    }
  }
}

Debug Output

No response

Expected Behavior

terraform plan should have failed, because terraform apply will fail

Actual Behavior

terraform plan passes, but terraform apply fails with the error: Error: Error creating Index: googleapi: Error 400: algorithmConfig is required but missing from the metadata.

Steps to reproduce

1. terraform plan
2. see that the plan succeeds
3. terraform apply
4. see that the apply fails due to algorithmConfig being required

Important Factoids

No response

References

No response

b/341993952

[ggtisc]

Hi @natedogith1 please provide the vars you are managing to replicate this issue (for sensitive data you can just put a comment like #provided)

[natedogith1]

I haven't tested, but I expect you'd get a similar error with:

resource "google_vertex_ai_index" "index" {
  display_name = "test-index"
}

[ggtisc]

This example was replicated successfully without errors in terraform plan and terraform apply. I suggest you follow the next examples because everything indicates that the variables that you have declared in your metadata configuration are causing this error.

[natedogith1]

Here's the error I get from terraform apply after adding each level of optional argument.

local.project_id is a project id, local.region is "europe-west3", local.contents_delta_gcs_uri is a gcs:// uri that points to an empty bucket.

resource "google_vertex_ai_index" "index" {
    project             = local.project_id
    display_name        = "test-display-name"
    index_update_method = "BATCH_UPDATE"
    region              = local.region
}

Error: Error creating Index: googleapi: Error 400: Index metadata is missing.

resource "google_vertex_ai_index" "index" {
    project             = local.project_id
    display_name        = "test-display-name"
    index_update_method = "BATCH_UPDATE"
    region              = local.region
    metadata {
        contents_delta_uri    = local.contents_delta_gcs_uri
        is_complete_overwrite = false
    }
}

Error: Error creating Index: googleapi: Error 400: dimensions is required but missing from Index metadata.

resource "google_vertex_ai_index" "index" {
    project             = local.project_id
    display_name        = "test-display-name"
    index_update_method = "BATCH_UPDATE"
    region              = local.region
    metadata {
        contents_delta_uri    = local.contents_delta_gcs_uri
        is_complete_overwrite = false
        config {
            dimensions                  = 3
        }
    }
}

Error: Error creating Index: googleapi: Error 400: algorithmConfig is required but missing from the metadata.

resource "google_vertex_ai_index" "index" {
    project             = local.project_id
    display_name        = "test-display-name"
    index_update_method = "BATCH_UPDATE"
    region              = local.region
    metadata {
        contents_delta_uri    = local.contents_delta_gcs_uri
        is_complete_overwrite = false
        config {
            dimensions                  = 3
            algorithm_config {
            }
        }
    }
}

Error: Error creating Index: googleapi: Error 400: algorithmConfig is required but missing from the metadata.

resource "google_vertex_ai_index" "index" {
    project             = local.project_id
    display_name        = "test-display-name"
    index_update_method = "BATCH_UPDATE"
    region              = local.region
    metadata {
        contents_delta_uri    = local.contents_delta_gcs_uri
        is_complete_overwrite = false
        config {
            dimensions                  = 3
            algorithm_config {
                brute_force_config {}
            }
        }
    }
}

Error: Error waiting to create Index: Error waiting for Creating Index: Error code 9, message: FAILED_PRECONDITION

[ggtisc]

Following the next configuration that is available on the shared link with your specifications everything was successfully replicated again without errors with the next configuration:

resource "google_storage_bucket" "bucket_17804" {
  name     = "vertex-ai-index-test-17804"
  location = "europe-west3"
  uniform_bucket_level_access = true
}

resource "google_vertex_ai_index" "google_vertex_ai_index_17804" {
    display_name        = "google-vertex-ai-index-17804"
    index_update_method = "BATCH_UPDATE"
    region              = "europe-west3"
    metadata {
        contents_delta_uri    = "gs://${google_storage_bucket.bucket_17804.name}/contents"
        is_complete_overwrite = false
        config {
            dimensions                  = 3
            algorithm_config {
                brute_force_config {}
            }
        }
    }
}

I suggest you check your user and account permissions and follow the documentation IO.

[ggtisc]

After a new replication with just this code the issue was confirmed:

resource "google_vertex_ai_index" "google_vertex_ai_index_17804" {
    display_name        = "google-vertex-ai-index-17804"
    index_update_method = "BATCH_UPDATE"
    region              = "europe-west3"
}

It is returning the next message even when the metadata appears as an optional argument in terraform registry:

Error: Error creating Index: googleapi: Error 400: Index metadata is missing.

[JoyceYingZhu]

metadata.contents_delta_uri is Required according to terraform registry