Skip to main content

Provision Infrastructure & Handle Drift

Use Infrastructure templates to provision and manage resources such as virtual machines, networks, and storage using Terraform. This guide walks through the deployment workflow, including planning changes, applying updates, and detecting configuration drift.

Create a deployment​

There are two primary ways to initiate a Terraform template deployment in GlobalAI:

  • From a Template: Navigate to Infrastructure > Terraform Template, select a template, and click Deploy template.
Deploy template button
  • From a the Terraform Deployments page: Navigate to Infrastructure > Terraform Deployment and click Create Deployment. This allows you to select a template from the list to start a new instance.
Create Deployment button

Deploy infrastructure​

Once you have initiated a deployment, configure the following details to provision your resources:

  1. Start completing the deployment metadata:
  • Name: Enter a unique name for your deployment.
  • Description: Provide a brief summary of the deployment's purpose.
  • Tags: Add optional tags to help organize and filter your deployments in the future.
  1. Configure the Deployment lifecycle. Manage how long your resources remain active by defining the Environment action and configuring timing settings:
  • Environment action: Select the primary lifecycle behavior for your deployment.
    • Manual: The deployment remains active until you delete it manually.
    • Terminate: The deployment is automatically removed after the configured Default time, plus any Grace time.
  • ETA (Expiration Time): Specify the exact date and time when the deployment should be terminated. While a default time is provided, you can customize this to fit your specific project needs
  • Grace time: Set an additional waiting period—either in minutes or ISO 8601 format (e.g., PT30M for 30 minutes)—that occurs after the ETA before the termination action is triggered.

  1. Configure the Input variables as required by the template. You can use values from the Vault for sensitive credentials.

  2. Select your deployment execution method:

    • Deploy & Plan: Generates a Terraform execution plan without making changes. Use this to preview exactly which resources will be created, modified, or removed in the Details console logs.
    • Deploy & Apply: Immediately starts the provisioning process based on the template configuration.
Configure Deployment form

Check for drift​

Configuration drift occurs when the actual state of your infrastructure differs from the configuration defined in the Terraform template. This typically happens when resources are modified outside of the GlobalAI platform.

To detect and reconcile drift:

  1. Navigate to the specific deployment you want to check.
  2. Click Check for drift.
  3. GlobalAI generates a new plan highlighting the differences between the live environment and the template.
  4. Review the plan in the console logs.
  5. If changes are detected, click Apply to reconcile the infrastructure and maintain consistency.
How can I verify changes before they are applied?

Always review the Plan output in the Details console. It shows a live-updated log of every resource action before the infrastructure is modified.

Check for drift button

Now that you know how to provision infrastructure and handle drift, you can explore these related guides.

Predefined Variables

Learn how predefined variables work in GlobalAI and how to use them to generate dynamic values in infrastructure configurations.

Add a Cluster on GlobalAI

Connect and configure the Kubernetes clusters where your infrastructure runs.

View Usage Metrics

Track cost and resource consumption for Terraform deployments.