Module for Vertex AI Workbench Instance

This module is used to create Vertex AI Workbench Instance. You can find example(s) for this module here

module "simple_vertex_ai_workbench" {
  source   = "GoogleCloudPlatform/vertex-ai/google//modules/workbench"
  version  = "~> 2.0"

  name       = "simple-vertex-ai-workbench"
  location   = "us-central1-a"
  project_id = var.project_id

  machine_type = "e2-standard-2"
  tags         = ["abc", "def"]

  labels = {
    env  = "test"
    type = "workbench"
  }

  service_accounts = [
    {
      email = google_service_account.workbench_sa.email
    },
  ]

  data_disks = [
    {
      disk_size_gb = 330
      disk_type    = "PD_BALANCED"
    },
  ]

  network_interfaces = [
    {
      network  = module.test-vpc-module.network_id
      subnet   = module.test-vpc-module.subnets_ids[0]
      nic_type = "GVNIC"
    }
  ]
}

Inputs

Name	Description	Type	Default	Required
accelerator_configs	The hardware accelerators used on this instance. If you use accelerators, make sure that your configuration has enough vCPUs and memory to support the machine_type you have selected. Currently supports only one accelerator configuration. Possible values for type: NVIDIA_TESLA_P100, NVIDIA_TESLA_V100, NVIDIA_TESLA_P4, NVIDIA_TESLA_T4, NVIDIA_TESLA_A100, NVIDIA_A100_80GB, NVIDIA_L4, NVIDIA_TESLA_T4_VWS, NVIDIA_TESLA_P100_VWS, NVIDIA_TESLA_P4_VWS	list(object({ type = optional(string) core_count = optional(number) }))	`null`	no
boot_disk_size_gb	The size of the boot disk in GB attached to this instance, up to a maximum of 64000 GB (64 TB). If not specified, this defaults to the recommended value of 150GB	`number`	`150`	no
boot_disk_type	Indicates the type of the boot disk. Possible values are: PD_STANDARD, PD_SSD, PD_BALANCED, PD_EXTREME	`string`	`"PD_BALANCED"`	no
confidential_instance_type	Defines the type of technology used by the confidential instance. Possible values are: SEV	`string`	`null`	no
container_image	Use a container image to start the workbench instance. repository path in format gcr.io/{project_id}/{imageName}. If tag is not specified, this defaults to the latest tag	object({ repository = optional(string) tag = optional(string) })	`null`	no
data_disks	Data disks attached to the VM instance. Currently supports only one data disk	list(object({ disk_size_gb = optional(number, 100) disk_type = optional(string, "PD_BALANCED") }))	`null`	no
desired_state	Desired state of the Workbench Instance. Set this field to ACTIVE to start the Instance, and STOPPED to stop the Instance	`string`	`null`	no
disable_proxy_access	If true, the workbench instance will not register with the proxy	`bool`	`false`	no
disable_public_ip	If true, no external IP will be assigned to this VM instance	`bool`	`true`	no
disk_encryption	Disk encryption method used on the boot and data disks, defaults to GMEK. Possible values are: GMEK, CMEK	`string`	`"GMEK"`	no
enable_ip_forwarding	Flag to enable ip forwarding or not, default false/off	`bool`	`false`	no
enable_third_party_identity	Flag that specifies that a notebook can be accessed with third party identity provider	`bool`	`null`	no
instance_id	User-defined unique ID of this instance	`string`	`null`	no
instance_owners	The owner of this instance after creation. Format: test@example.com Currently supports one owner only. If not specified, all of the service account users of your VM instance''s service account can use the instance	`list(string)`	`[]`	no
kms_key	The KMS key used to encrypt the disks, only applicable if disk_encryption is CMEK. Format: projects/{project_id}/locations/{location}/keyRings/{key_ring_id}/cryptoKeys/{key_id}	`string`	`null`	no
labels	Labels to apply to this instance	`map(string)`	`{}`	no
location	Zone in which workbench instance should be created	`string`	n/a	yes
machine_type	The machine type of the VM instance	`string`	`null`	no
metadata	Custom metadata to apply to this instance	`map(string)`	`{}`	no
metadata_configs	predefined metadata to apply to this instance	object({ idle-timeout-seconds = optional(number) notebook-upgrade-schedule = optional(string) notebook-disable-downloads = optional(bool) notebook-disable-root = optional(bool) post-startup-script = optional(string) post-startup-script-behavior = optional(string) nbconvert = optional(bool) notebook-enable-delete-to-trash = optional(bool) disable-mixer = optional(bool) jupyter-user = optional(string) report-event-health = optional(bool) enable-guest-attributes = optional(bool) serial-port-logging-enable = optional(bool) container-allow-fuse = optional(bool) install-unattended-upgrades = optional(bool) })	`{}`	no
name	The name of this workbench instance	`string`	n/a	yes
network_interfaces	The network interfaces for the VM. Supports only one interface. The nic_type of vNIC to be used on this interface. This may be gVNIC or VirtioNet. Possible values are: VIRTIO_NET, GVNIC	list(object({ network = optional(string) nic_type = optional(string) subnet = optional(string) }))	`null`	no
project_id	The ID of the project in which the resource belongs	`string`	n/a	yes
service_accounts	The service account that serves as an identity for the VM instance. Currently supports only one service account	list(object({ email = optional(string) }))	`null`	no
shielded_instance_config	A set of Shielded Instance options	object({ enable_secure_boot = optional(bool, false) enable_vtpm = optional(bool, true) enable_integrity_monitoring = optional(bool, true) })	`null`	no
tags	The Compute Engine tags to add to instance	`list(string)`	`null`	no
vm_image	Definition of a custom Compute Engine virtual machine image for starting a workbench instance with the environment installed directly on the VM	object({ family = optional(string) name = optional(string) project = optional(string) })	`null`	no

Outputs

Name	Description
creator	Email address of entity that sent original CreateInstance request.
health_info	Additional information about the the Vertex AI Workbench instance's health.
health_state	The health state of the Vertex AI Workbench instance.
id	The Vertex AI Workbench instance ID.
proxy_uri	The proxy endpoint that is used to access the Jupyter notebook.
state	The state of the Vertex AI Workbench instance.
upgrade_history	The upgrade history of the Vertex AI Workbench instance.
work_bench	Workbenchs created
workbench_name	The name of the Vertex AI Workbench instance

Manage features through metadata

Values for metadata_configs parameters are available here

Requirements

These sections describe requirements for using this module.

Software

The following dependencies must be available:

[Terraform][terraform] v1.3+
[Terraform Provider for GCP][terraform-provider-gcp] plugin v6.42+

Enable API's

In order to operate with the Service Account you must activate the following API on the project where the Service Account was created:

compute.googleapis.com
iap.googleapis.com
aiplatform.googleapis.com
notebooks.googleapis.com
cloudkms.googleapis.com

Service Account

A service account with the following roles must be used to provision the resources of this module:

Notebooks Admin: roles/notebooks.admin
Cloud KMS CryptoKey Encrypter/Decrypter : roles/cloudkms.cryptoKeyEncrypterDecrypter
Cloud KMS Admin: roles/cloudkms.admin
Compute Network Admin: roles/compute.networkAdmin
Project IAM Admin: roles/resourcemanager.projectIamAdmin
Service Account Admin: roles/iam.serviceAccountAdmin
Service Account User: roles/iam.serviceAccountUser