CDKTF Google Provider

16.12.1 verified Fri May 01 auth: no python

CDKTF (CDK for Terraform) prebuilt bindings for the Google Cloud Platform (GCP) provider. Version 16.12.1 targets Terraform provider google ~> 6.0. Released frequently, follows cdktf provider releases.

pip install cdktf-cdktf-provider-google

Common errors

error ModuleNotFoundError: No module named 'cdktf_cdktf_provider_google' ↓

cause Package not installed or imported incorrectly.

fix

Run pip install cdktf-cdktf-provider-google and import as from cdktf_cdktf_provider_google import google.

error AttributeError: module 'cdktf_cdktf_provider_google' has no attribute 'ComputeInstance' ↓

cause Direct import of resource without submodule.

fix

Import from submodule: from cdktf_cdktf_provider_google.compute_instance import ComputeInstance.

error TypeError: __init__() takes at least 4 arguments (3 given) ↓

cause Missing or wrong number of positional arguments; often missing construct `id`.

fix

Ensure the second argument is the construct ID string. Example: ComputeInstance(stack, 'my_instance', ...).

error Error: Unsupported attribute: "machineType" ↓

cause Using camelCase property name instead of snake_case.

fix

Replace machineType with machine_type (and similarly for all properties).

error Error: Invalid provider configuration on main.tf.json line X: provider.google: missing required field 'project' ↓

cause GoogleProvider not configured properly.

fix

Add GoogleProvider to stack with required fields: google.GoogleProvider(stack, 'google', project='...').

Warnings

breaking Version 16.x drops support for Python 3.8 (requires ~=3.9). ↓

fix Upgrade to Python 3.9+.

breaking Breaking changes between major provider versions (e.g., from 5.x to 6.x) reflect upstream Terraform provider changes. Always pin provider version and run `cdktf diff` after upgrades. ↓

fix Check upstream Terraform google provider changelog and update resource configs accordingly.

gotcha Property names use snake_case in Python, not camelCase as in HCL. Common mistake: `machineType` instead of `machine_type`. ↓

fix Use snake_case for all resource properties. Refer to CDKTF documentation for bindings.

gotcha Resource constructor expects `id` as second positional argument (the construct ID). Omitting it or reordering arguments causes cryptic errors. ↓

fix Always provide `id` argument: `Resource(stack, 'id', ...)`.

Imports

google
wrong
```
import cdktf_cdktf_provider_google
```
correct
```
from cdktf_cdktf_provider_google import google
```
The provider is imported as a module named 'google', not the pypi package name.

ComputeInstance

wrong

from cdktf_cdktf_provider_google import ComputeInstance

correct

from cdktf_cdktf_provider_google.compute_instance import ComputeInstance

Resources are in submodules; direct import fails.

DataGoogleComputeInstance

wrong

from cdktf_cdktf_provider_google import DataGoogleComputeInstance

correct

from cdktf_cdktf_provider_google.data_google_compute_instance import DataGoogleComputeInstance

Data sources also in submodules.

Quickstart

Minimal stack creating a GCE instance and synthesizing Terraform JSON.

from cdktf import App, TerraformStack
from cdktf_cdktf_provider_google import google
from cdktf_cdktf_provider_google.compute_instance import ComputeInstance

app = App()
stack = TerraformStack(app, "quickstart")
google.GoogleProvider(stack, "provider", project="my-project", region="us-central1")
ComputeInstance(stack, "vm", name="test", machine_type="e2-micro", zone="us-central1-a", boot_disk={"initialize_params": {"image": "debian-cloud/debian-11"}}, network_interface=[{"network": "default"}])
app.synth()