terraform-google-cloud-function

A Terraform module for deploying Google Cloud Functions Gen 2 triggered by Pub/Sub messages with automatic source code packaging and deployment.

Features

• Automatic source code zipping and upload to GCS
• Cloud Function Gen 2 deployment with Pub/Sub trigger
• Lifecycle management for old function archives
• Configurable runtime, memory, timeout, and scaling
• Environment variable support
• Built-in retry policy configuration

---

Assumptions

The project assumes the following:

• A basic understanding of Git.
• Git version >= 2.33.0.
• An existing GCP IAM user or role with access to create/update/delete resources defined in main.tf.
• GCloud CLI >= 465.0.0.
• A basic understanding of Terraform.
• Terraform version >= 1.3.0.
• (Optional - for local testing) A basic understanding of Make.
  • Make version >= GNU Make 3.81.
  • Important Note: This project includes a Makefile to speed up local development in Terraform. The make targets act as a wrapper around Terraform commands. As such, make has only been tested/verified on Linux/Mac OS. Though, it is possible to install make using Chocolatey, we do not guarantee this approach as it has not been tested/verified. You may use the commands in the Makefile as a guide to run each Terraform command locally on Windows.

---

Test

Important Note: This project includes a Makefile to speed up local development in Terraform. The make targets act as a wrapper around Terraform commands. As such, make has only been tested/verified on Linux/Mac OS. Though, it is possible to install make using Chocolatey, we do not guarantee this approach as it has not been tested/verified. You may use the commands in the Makefile as a guide to run each Terraform command locally on Windows.

gcloud init # https://cloud.google.com/docs/authentication/gcloud
gcloud auth application-default login

# Copy the example tfvars and customize it
cp examples/simple/examples.tfvars examples/simple/terraform.tfvars
# Edit terraform.tfvars with your values

# Run terraform commands
make plan
make apply
make destroy

---

Contributions

Contributions are always welcome. As such, this project uses the main branch as the source of truth to track changes.

Step 1. Clone this project.

# Using SSH
$ git clone git@github.com:nurdsoft/terraform-google-cloud-function.git

# Using HTTPS
$ git clone https://github.com/nurdsoft/terraform-google-cloud-function.git

Step 2. Checkout a feature branch: git checkout -b feature/abc.

Step 3. Validate the change/s locally by executing the steps defined under Test.

Step 4. If testing is successful, commit and push the new change/s to the remote.

$ git add file1 file2 ...

$ git commit -m "Adding some change"

$ git push --set-upstream origin feature/abc

Step 5. Once pushed, create a PR and assign it to a member for review.

• Important Note: It can be helpful to attach the terraform plan output in the PR.

Step 6. A team member reviews/approves/merges the change/s.

Step 7. Once merged, deploy the required changes as needed.

Step 8. Once deployed, verify that the changes have been deployed.

• If possible, please add a plan output using the feature branch so the member reviewing the PR has better visibility into the changes.

---

Usage

module "cloud_function" {
  source = "git::https://github.com/nurdsoft/terraform-google-cloud-function.git?ref=main"

  project_id           = "my-project"
  region               = "us-central1"
  function_source_dir  = "./my-function"
  function_entry_point = "HandleMessage"
  pubsub_topic_id      = "projects/my-project/topics/my-topic"

  environment_variables = {
    SLACK_WEBHOOK_URL = var.slack_webhook_url
  }
}

Examples

Example	Description
simple	Minimal setup with a sample Go function

Requirements

Name	Version
terraform	>= 1.3
google	>= 5.0
archive	>= 2.0

Providers

Name	Version
google	>= 5.0
archive	>= 2.0

Inputs

Required

Name	Description	Type	Default	Required
project_id	The GCP project ID to deploy resources into	string	n/a	yes
region	The GCP region to deploy the Cloud Function and Storage Bucket	string	n/a	yes
function_source_dir	The local directory path containing the Cloud Function source code to be zipped and deployed	string	n/a	yes
function_entry_point	The name of the function to execute when the Cloud Function is triggered	string	n/a	yes
pubsub_topic_id	The fully-qualified Pub/Sub topic ID that triggers the Cloud Function (e.g. projects/my-project/topics/my-topic)	string	n/a	yes

Optional

Name	Description	Type	Default	Required
function_bucket_name	The base name of the GCS bucket used to store the function source code. The project_id will be appended to ensure global uniqueness	string	"cloud-function-source"	no
bucket_lifecycle_age_days	Number of days after which old function archives are automatically deleted from the bucket	number	30	no
function_name	The name of the Cloud Function	string	"pubsub-triggered-function"	no
function_description	The description of the Cloud Function	string	"Cloud Function triggered by Pub/Sub messages"	no
function_runtime	The runtime environment for the Cloud Function (e.g. go125, python311, nodejs20)	string	"go125"	no
max_instance_count	The maximum number of function instances that can run simultaneously	number	3	no
available_memory	The amount of memory available for the function (e.g. 256Mi, 512Mi, 1Gi)	string	"512Mi"	no
timeout_seconds	The maximum amount of time the function can run before timing out (in seconds)	number	60	no
environment_variables	A map of environment variables to pass to the Cloud Function	map(string)	{}	no
event_trigger_type	The type of event that triggers the function	string	"google.cloud.pubsub.topic.v1.messagePublished"	no
retry_policy	The retry policy for failed function executions (RETRY_POLICY_RETRY or RETRY_POLICY_DO_NOT_RETRY)	string	"RETRY_POLICY_DO_NOT_RETRY"	no

Outputs

Name	Description
function_id	The fully-qualified ID of the Cloud Function
function_name	The name of the Cloud Function
function_url	The URL of the Cloud Function service
bucket_name	The name of the GCS bucket storing the function source code
bucket_url	The URL of the GCS bucket

Authors

Module is maintained by Nurdsoft.

License

Apache 2 Licensed. See LICENSE for full details.