Terraform GCP

You can deploy CubeAPM on GCP (Compute Engine VM) using our Terraform code.

Prerequisites

Before getting started, ensure you have the following configuration ready:

• Terraform installed (v1.0.0 or higher).
• Google Cloud Platform (GCP) credentials configured locally (e.g., via gcloud auth application-default login).
• An existing GCP VPC and Subnetwork where CubeAPM will be deployed.

Installation

1. Clone the CubeAPM-Terraform-VM-GCP repository:

   git clone https://github.com/cubeapm/CubeAPM-Terraform-VM-GCP.git

2. Initialize Terraform:

   cd CubeAPM-Terraform-VM-GCP
   terraform init

3. Deploy CubeAPM:

   While applying terraform it asks for following values, provide as per your requirement:

   Variable	Type	Default	Description
   gcp_project_id	string		The GCP Project ID where resources will be deployed.
   gcp_region	string	"asia-south1"	The GCP Region.
   gcp_zone	string	"asia-south1-a"	The GCP zone to deploy the resources to.
   gcp_image	string	"ubuntu-os-cloud/ubuntu-2404-lts-amd64"	The GCP image to use for the instance.
   gcp_machine_type	string	"e2-standard-4"	The GCP machine type to use.
   gcp_network	string		VPC network name or self-link.
   gcp_subnetwork	string		Subnetwork name or self-link.
   load_balancer_scope	string		Select the load balancer scope: "global" or "regional".
   load_balancer_type	string		Select the load balancer type: "internal" or "internet-facing".
   create_lb	boolean	true	Set to true to create a new Load Balancer (URL Map, Target Proxy, Forwarding Rule). If false, only the Backend Service is created.

   warning

   To avoid being prompted for these variables every time you run Terraform, create a terraform.tfvars file in the same directory:

4. Here's a sample terraform.tfvars file you can take a reference from here and update your terraform.tfvars file based on your infrastructure requirements:

   Creating new LB

   Example 1: Creating a New Load Balancer

   # GCP Authentication & Location
   gcp_project_id        = "my-cubeapm-project-12345"
   gcp_region            = "asia-south1"
   gcp_zone              = "asia-south1-a"
   
   # VM Specifications
   gcp_image             = "ubuntu-os-cloud/ubuntu-2404-lts-amd64"
   gcp_machine_type      = "e2-standard-4"
   
   # Networking
   gcp_network           = "my-vpc-network"
   gcp_subnetwork        = "my-app-subnet"
   
   # Load Balancer Configuration
   create_lb             = true
   load_balancer_scope   = "global"
   load_balancer_type    = "internet-facing"

   Using existing LB

   Example 2: Using an Existing Load Balancer

   # GCP Authentication & Location
   gcp_project_id        = "my-cubeapm-project-12345"
   gcp_region            = "asia-south1"
   gcp_zone              = "asia-south1-a"
   
   # VM Specifications
   gcp_image             = "ubuntu-os-cloud/ubuntu-2404-lts-amd64"
   gcp_machine_type      = "e2-standard-4"
   
   # Networking
   gcp_network           = "my-vpc-network"
   gcp_subnetwork        = "my-app-subnet"
   
   # Load Balancer Configuration
   create_lb             = false
   load_balancer_scope   = "global"
   load_balancer_type    = "internet-facing"
   info

   Why are load_balancer_scope and load_balancer_type required when create_lb = false? Even when you are using an existing Load Balancer, Terraform still needs to create a Backend Service for you to attach to your URL Map. In GCP, Backend Services are categorized by load balancer type. Terraform needs these variables to know whether to provision a Global or Regional backend service, and whether to configure it for Internal or External traffic, ensuring it is compatible with your existing Load Balancer.

5. Once your variables are configured, run the following command to provision the infrastructure:

   # use this syntax if you have created `terraform.tfvars` file.
   terraform apply
   
   # use this syntax if you don't want to create the `terraform.tfvars` file.
   terraform apply -var="variable_name=variable_value" ...

Knowledge Base

For more information about Terraform for GCP, please refer to the Terraform GCP documentation.