☁️ Google Cloud Vertex AI Index Endpoint Deployed Index Terraform Module

Deploys a Vertex AI Vector Search Index onto an Index Endpoint for online nearest-neighbor query serving, targeting hashicorp/google ~> 7.0.

🧩 Overview

Creates exactly one google_vertex_ai_index_endpoint_deployed_index resource — this module performs the actual deployment step terraform-google-vertex-ai-index-endpoint's own README explicitly said was out of scope.
This is the first module in this catalog with TWO required parent-scoping cross-module inputs (index + index_endpoint) — every other Vertex AI module so far has had at most one.
No labels field exists on this resource at all — the first module in this domain's slice to genuinely lack labels support.
Confirmed timeout asymmetry: create/update default to 45 minutes, delete to 20 minutes — longer than this domain's usual uniform 20-minute pattern, reflecting genuinely slow deployment operations.
enable_access_logging defaults to true — a newly-documented extension of this suite's general audit-logging secure-default row to this Vertex AI module.

💡 Why it matters: Building an Index and an Index Endpoint doesn't serve any queries on its own — this module is the missing link that actually puts an Index online for nearest-neighbor lookups.

❤️ Support this project

If these Terraform modules have been helpful to you or your organization, I'd appreciate your support in any of the following ways:

⭐ Star this repository to help others discover this Terraform module.
🤝 Connect with me on LinkedIn: linkedin.com/in/microsoftexpert
☕ Buy me a coffee: buymeacoffee.com/microsoftexpert

Whether it's a star, a professional connection, or a coffee, every gesture helps keep these modules actively maintained and continually improving. Thank you for being part of the community!

🗺️ Where this fits

flowchart LR
 IDX["terraform-google-vertex-ai-index"]:::keystone
 IEP["terraform-google-vertex-ai-index-endpoint"]:::keystone
 THIS["terraform-google-vertex-ai-index-endpoint-deployed-index<br/>(this module)"]:::thismodule
 GA["terraform-google-global-address<br/>(optional)"]:::external
 SA["terraform-google-service-account<br/>(optional)"]:::external

 IDX -->|"id -> index (required)"| THIS
 IEP -->|"id -> index_endpoint (required)"| THIS
 GA -.->|"name -> reserved_ip_ranges (optional)"| THIS
 SA -.->|"email -> deployed_index_auth_config.auth_provider.allowed_issuers (optional)"| THIS

 classDef thismodule fill:#4285F4,color:#FFFFFF,stroke:#174EA6,stroke-width:1px;
 classDef keystone fill:#174EA6,color:#FFFFFF,stroke:#174EA6,stroke-width:1px;
 classDef external fill:#E8EAED,color:#202124,stroke:#9AA0A6,stroke-width:1px,stroke-dasharray: 3 3;

Validated via the Mermaid Chart MCP (valid: true). Part of the existing data-ml subgraph. Both terraform-google-vertex-ai-index and terraform-google-vertex-ai-index-endpoint are equally required upstream keystones — neither is more load-bearing than the other.

🧬 What this builds

flowchart TB
 subgraph Inputs
 DEPID["var.deployed_index_id"]
 IDX["var.index"]
 IEP["var.index_endpoint"]
 AUTOR["var.automatic_resources"]
 DEDR["var.dedicated_resources"]
 end

 KEYSTONE["google_vertex_ai_index_endpoint_deployed_index.this<br/>(two required parents)"]:::keystone

 subgraph Outputs
 ID["id"]
 NM["name"]
 CT["create_time"]
 end

 DEPID --> KEYSTONE
 IDX --> KEYSTONE
 IEP --> KEYSTONE
 AUTOR --> KEYSTONE
 DEDR --> KEYSTONE

 KEYSTONE --> ID
 KEYSTONE --> NM
 KEYSTONE --> CT

 classDef keystone fill:#174EA6,color:#FFFFFF,stroke:#174EA6,stroke-width:1px;

Validated via the Mermaid Chart MCP (valid: true). Visually distinguish this from every prior single-parent Vertex AI module in this catalog — two required parent inputs feed the keystone.

Resource inventory: google_vertex_ai_index_endpoint_deployed_index.this (1 resource).

✅ Provider / Versions

Component	Requirement
Terraform	`>= 1.12.0`
`hashicorp/google`	`~> 7.0` (resolved v7.39.0 during authoring)
Provider block	None — the caller configures `google`

Schema notes that bite:

deployed_indexes does not exist on the Index Endpoint resource itself — this module is the real, separately Terraform-manageable deployment mechanism, background context for why it exists at all.
No labels field at all.
Confirmed 45/45/20 timeout asymmetry — not this domain's usual uniform 20 minutes.
No self_link and no update_time attribute — the missing update_time is a genuine asymmetry with every other Vertex AI module in this catalog.
automatic_resources/dedicated_resources are NOT textually confirmed mutually exclusive — unlike the sibling Index Endpoint's network/private_service_connect_config pair.

🔑 Required IAM Roles

roles/aiplatform.user on the target project — deploy/undeploy indexes; same role as every other module in this domain.

☁️ GCP Prerequisites

aiplatform.googleapis.com enabled (via terraform-google-project-services).
Both the target Index and the target Index Endpoint must already exist and be applied first — hard dependencies, not optional cross-references.
If reserved_ip_ranges is used, the referenced GCE address(es) must already exist under the peered VPC.
Deployment operations are genuinely slow — expect apply times consistent with the confirmed 45-minute create/update default.

📁 Module Structure

terraform-google-vertex-ai-index-endpoint-deployed-index/
├── providers.tf
├── variables.tf # deployed_index_id, index, index_endpoint, automatic_resources,...
├── main.tf # google_vertex_ai_index_endpoint_deployed_index.this
├── outputs.tf # id, name, create_time, index_sync_time, private_endpoints — no self_link/update_time/labels
├── README.md
├── SCOPE.md
└── examples/
 └── basic/

⚙️ Quick Start

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "prodDeployment01"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
}

🔌 Cross-Module Contract

Consumes:

Input	Type	Source module
`index`	`string` (`.id` form)	`terraform-google-vertex-ai-index` (required)
`index_endpoint`	`string` (`.id` form)	`terraform-google-vertex-ai-index-endpoint` (required)
`reserved_ip_ranges` entries	`string` (`.name` form)	`terraform-google-global-address` (optional)
`deployed_index_auth_config.auth_provider.allowed_issuers` entries	`string` (`.email` form)	`terraform-google-service-account` (optional)

Emits:

Output	Description	Consumed by
`id`	Composite `{{index_endpoint}}/deployedIndex/{{deployed_index_id}}` — no self_link	None — terminal
`name`	Computed	None
`create_time`	No corresponding `update_time`	None
`index_sync_time`	Latest-state sync flag	None
`private_endpoints`	Conditional/informational	None

📚 Example Library

1 · Minimal deployment — three required fields only

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "basicDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
}

ℹ️ automatic_resources/dedicated_resources left entirely at provider defaults here.

2 · automatic_resources with explicit replica counts

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "autoScaledDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id

  automatic_resources = {
    min_replica_count = 2
    max_replica_count = 10
  }
}

3 · dedicated_resources with an explicit machine_type

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "dedicatedDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id

  dedicated_resources = {
    machine_spec = {
      machine_type = "e2-standard-16"
    }
    min_replica_count = 2
    max_replica_count = 4
  }
}

🔒 No machine_type/replica-count default is baked into this module — the caller must make an explicit, informed sizing choice, mirroring the cost-safety framing already established for terraform-google-vertex-ai-deployment-resource-pool.

4 · deployed_index_auth_config wired to a service account

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "authenticatedDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id

  deployed_index_auth_config = {
    auth_provider = {
      allowed_issuers = [module.query_client_sa.email]
      audiences       = ["https://query-client.example.com"]
    }
  }
}

5 · reserved_ip_ranges + deployment_group

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id  = "prodDeployment"
  index              = module.index.id
  index_endpoint     = module.index_endpoint.id
  reserved_ip_ranges = [module.reserved_address.name]
  deployment_group   = "prod"
}

6 · enable_access_logging toggled off

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id     = "noLoggingDeployment"
  index                 = module.index.id
  index_endpoint        = module.index_endpoint.id
  enable_access_logging = false
}

⚠️ This module defaults enable_access_logging to true per this suite's general audit-logging row — only disable it with an explicit, deliberate reason.

7 · A display_name

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "labeledDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
  display_name      = "Production Vector Search Deployment"
}

8 · Explicit region

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "regionalDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
  region            = "us-central1"
}

9 · Relying on the default deletion_policy = "PREVENT"

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "protectedDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
}

ℹ️ No deletion_policy set here — the module default of "PREVENT" applies. Destroying a DeployedIndex that is actively serving production query traffic is a real operational hazard.

10 · deletion_policy = "DELETE" opt-out

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "scratchDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
  deletion_policy   = "DELETE"
}

11 · deletion_policy = "ABANDON"

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "migratingDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id
  deletion_policy   = "ABANDON"
}

12 · Custom timeouts (45/45/20 default, not 20/20/20)

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "extendedTimeoutDeployment"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id

  timeouts = {
    create = "90m"
    update = "90m"
    delete = "30m"
  }
}

ℹ️ This resource's own provider defaults are already asymmetric — 45 minutes for create/update, 20 minutes for delete — do not assume a uniform 20-minute baseline when overriding.

13 · 🏗️ End-to-end composition

Completing the three-module Index → Index Endpoint → Deployed Index chain this catalog has now fully closed.

module "index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index.git?ref=v1.0.0"

  display_name = "casey-prod-vector-index"
  region       = "us-central1"
  #... metadata, index_update_method, etc.
}

module "index_endpoint" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint.git?ref=v1.0.0"

  display_name = "casey-prod-index-endpoint"
  region       = "us-central1"
  #... network or public_endpoint_enabled, etc.
}

module "deployed_index" {
  source = "git::https://github.com/microsoftexpert/terraform-google-vertex-ai-index-endpoint-deployed-index.git?ref=v1.0.0"

  deployed_index_id = "prodDeployment01"
  index             = module.index.id
  index_endpoint    = module.index_endpoint.id

  automatic_resources = {
    min_replica_count = 2
    max_replica_count = 10
  }
}

📥 Inputs

Variable	Type	Default	Notes
`deployed_index_id`	`string`	— (required)	Up to 128 chars, must start with a letter
`index`	`string`	— (required)	Consumes `terraform-google-vertex-ai-index`'s `id`
`index_endpoint`	`string`	— (required)	Consumes `terraform-google-vertex-ai-index-endpoint`'s `id`
`display_name`	`string`	`null`	Up to 128 UTF-8 chars
`automatic_resources`	`object({...})`	`null`	Not confirmed mutually exclusive with `dedicated_resources`
`dedicated_resources`	`object({...})`	`null`	No baked-in machine_type/replica-count default
`enable_access_logging`	`bool`	`true`	House secure-by-default extension
`deployed_index_auth_config`	`object({...})`	`null`	No default audiences/issuers
`reserved_ip_ranges`	`list(string)`	`[]`	Names, not ids/self_links
`deployment_group`	`string`	`null`	Up to 64 chars
`region`	`string`	`null`	No hardcoded allow-list
`deletion_policy`	`string`	`"PREVENT"`	Closed enum; house extension over provider default
`timeouts`	`object({...})`	`null`	create/update default 45 min, delete default 20 min

No labels variable — confirmed absent from this resource's schema entirely.

🧾 Outputs

Output	Description
`id`	Composite `{{index_endpoint}}/deployedIndex/{{deployed_index_id}}`
`name`	Computed
`create_time`	No corresponding `update_time` exists
`index_sync_time`	Latest-state sync flag
`private_endpoints`	Conditional/informational

No self_link, no update_time, no labels-related output.

🧠 Architecture Notes

Two required parent-scoping cross-module inputs — the single most structurally distinctive fact about this module in this catalog.
Confirmed 45/45/20 timeout asymmetry — index deployment provisions real serving infrastructure; expect apply times to reflect that.
Update-in-place finding: the schema's distinct update timeout (matching create, not delete) is real evidence of a genuine in-place update path (likely MutateDeployedIndex) — but deployed_index_id/index/index_endpoint are near-certainly force-new in practice as identity fields, an inference from their role, not a textually confirmed ForceNew flag.
No labels field — the first module in this domain's slice to lack it.

🧱 Design Principles

Concern	Secure default	Opt-out (explicit)
Access log visibility	`enable_access_logging = true` — house extension of this suite's general audit-logging row, applied to this Vertex AI module for the first time	Caller sets `false` explicitly
Accidental destroy of a live, serving deployment	`deletion_policy = "PREVENT"` (house extension over the provider's native `"DELETE"` default)	Caller sets `"DELETE"` or `"ABANDON"` explicitly
Compute sizing cost	No baked-in `automatic_resources`/`dedicated_resources` default — the caller must make an explicit, informed sizing choice, mirroring `terraform-google-vertex-ai-deployment-resource-pool`'s own cost-safety framing	N/A — deliberate absence of a default

🚀 Runbook

cd terraform-google-vertex-ai-index-endpoint-deployed-index
terraform init -backend=false
terraform validate
terraform fmt -check

🧪 Testing

validate/fmt cannot catch a "parent Index or Index Endpoint doesn't exist yet" apply-time failure, a genuinely long deployment exceeding the 45-minute default under load, or a reserved_ip_ranges address that doesn't actually exist — only a real plan/apply against a live project would.

💬 Example Output

$ terraform output
id = "projects/casey-prod/locations/us-central1/indexEndpoints/casey-prod-index-endpoint/deployedIndex/prodDeployment01"
name = "prodDeployment01"
create_time = "2026-07-14T12:00:00Z"
index_sync_time = "2026-07-14T12:05:00Z"

🔍 Troubleshooting

Symptom	Cause	Fix
Apply fails because the referenced Index or Index Endpoint does not exist	This module was applied before either parent resource	Apply `terraform-google-vertex-ai-index` and `terraform-google-vertex-ai-index-endpoint` first
Apply appears to hang past 45 minutes during index deployment	Index deployment is a genuinely slow API operation; large indexes or high load can exceed the default timeout	Increase `timeouts.create`/`timeouts.update` explicitly; this is expected behavior, not a bug
No `labels` input exists, unlike every other module in this domain	This resource's schema genuinely has no labels field — confirmed absent	There is no workaround; label the parent Index/Index Endpoint instead if tagging is required

🔗 Related Docs

google_vertex_ai_index_endpoint_deployed_index provider docs
terraform-google-vertex-ai-index-endpoint — this module's own SCOPE.md documented the gap this module closes
terraform-google-vertex-ai-index
This module's SCOPE.md

Navigation Menu

Repository navigation

About

Repository files navigation

☁️ Google Cloud Vertex AI Index Endpoint Deployed Index Terraform Module

🧩 Overview

❤️ Support this project

🗺️ Where this fits

🧬 What this builds

✅ Provider / Versions

🔑 Required IAM Roles

☁️ GCP Prerequisites

📁 Module Structure

⚙️ Quick Start

🔌 Cross-Module Contract

📚 Example Library

📥 Inputs

🧾 Outputs

🧠 Architecture Notes

🧱 Design Principles

🚀 Runbook

🧪 Testing

💬 Example Output

🔍 Troubleshooting

🔗 Related Docs

About

Releases

Packages

Contributors

Languages

Name		Name	Last commit message	Last commit date
examples/basic		examples/basic
.gitignore		.gitignore
LICENSE		LICENSE
README.md		README.md
SCOPE.md		SCOPE.md
main.tf		main.tf
outputs.tf		outputs.tf
providers.tf		providers.tf
variables.tf		variables.tf

Latest commit

History

Folders and files

Repository files navigation

☁️ Google Cloud Vertex AI Index Endpoint Deployed Index Terraform Module

🧩 Overview

❤️ Support this project

🗺️ Where this fits

🧬 What this builds

✅ Provider / Versions

🔑 Required IAM Roles

☁️ GCP Prerequisites

📁 Module Structure

⚙️ Quick Start

🔌 Cross-Module Contract

📚 Example Library

📥 Inputs

🧾 Outputs

🧠 Architecture Notes

🧱 Design Principles

🚀 Runbook

🧪 Testing

💬 Example Output

🔍 Troubleshooting

🔗 Related Docs

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages