Stacks overview
As your infrastructure grows, managing Terraform configurations becomes increasingly complex. Stacks are a powerful configuration layer in HCP Terraform that simplifies managing your infrastructure modules and then repeating that infrastructure.
Hands-on: Try out the Deploy a Stack with HCP Terraform tutorial to get started with Stacks quickly.
Stacks replace Terraform's traditional root module structure with a new configuration layer of modular components built on top of your Terraform child and shared modules. Stacks enable you to provision and coordinate your infrastructure lifecycle at scale, offering an organized and reusable approach that expands upon infrastructure as code (IaC).
Public Beta: Stacks are in public beta. Once we generally release them, we do not guarantee backward compatibility support for the public beta. All APIs, workflows, and HCL syntax for Stacks are subject to change.
Background
Stacks are an alternative way to organize your infrastructure and fundamentally differ from HCP Terraform workspaces. Stacks are not built on top of HCP Terraform workspaces, but can exist alongside them in the same project.
Previously, those looking to deploy repeatable infrastructure using HCP Terraform would create separate workspaces and then use run triggers or other automation tools to coordinate changes between them. However, workspaces are not truly coupled to each other, hampering your ability to flexibly manage infrastructure as it scales.
Terraform Stacks enable you to split your Terraform configuration into components and then deploy and manage those components across multiple environments. You can manage the lifecycle of each deployment separately, roll out configuration changes across your deployments, and manage your Stack as a unit in HCP Terraform.
Each deployment in a Stack represents a group of components that work together, such as a development and production environment. The configuration for each Stack is identical across deployments, and you can further customize each environment with input variables. You can then use HCP Terraform to roll out changes to each deployment and keep track of changes across your environments.
Stacks are unavailable for users on legacy HCP Terraform team plans. Learn more about migrating to current HCP Terraform plan.
Workflow
Start by creating a Stack configuration file and filling it with the component
blocks. Each component
block includes a Terraform module as its source, and you can configure a component
further using input arguments. Your Stack components share a lifecycle, which you can repeatedly deploy together using HCP Terraform.
After configuring your Stack’s components
, you create a separate deployment configuration file to define how you want to repeat its infrastructure. Each deployment in a Stack represents a group of infrastructure that works together. You can define deployment
blocks for your development environments, cloud provider accounts, or regions.
Once ready to deploy your Stack’s infrastructure, you can help manage your deployments by defining orchestration rules. For example, you can automatically approve operations whenever a deployment meets certain criteria.
Primary workflow
The overall process for managing your infrastructure using Stacks consists of the following steps.
Design a Stack
Consider your current infrastructure lifecycle and how the pieces interact with each other. Then, determine how to break down your infrastructure into components and think about how you want to deploy the infrastructure your Stack describes. Refer to Design a Stack to learn more.
Write a Stack configuration
Configure and define your Stack in a tfstack.hcl
file. To define a Stack, add and configure component
blocks, each representing an individual Terraform module. Refer to Define configuration to learn more.
Stacks declare providers in a different workflow than in traditional Terraform configurations, refer to Declare providers to learn more. Stacks also have their own way of authenticating with providers, refer to Authenticate a Stack for details.
Define deployments
With your Stack configuration file complete, the next step is to define how you want to deploy your Stack. Create a tfdeploy.hcl
file defining how HCP Terraform should deploy your Stack’s infrastructure. In Stacks, deployments allow you to replicate infrastructure across multiple environments, regions, or accounts. Each deployment runs in an isolated agent, ensuring that changes in one deployment do not affect others. Refer to Define deployment configuration to learn more.
Manage and orchestrate deployments
Use your Stack’s context to define orchestration rules with the orchestrate
block. You can define rules for automated approvals, error handling, and other events within your deployment configuration. Refer to Set conditions for deployment plans to learn more.
Deploy
After fully configuring your Stack, use HCP Terraform to deploy your Stack’s infrastructure. Refer to Create a Stack to learn more.
Guidance
Stacks are useful when managing complex, multi-environment infrastructures where consistency and reusability are crucial. Refer to Stack use cases for inspiration on situations where Stacks are particularly well suited.
Constraints, limitations, and troubleshooting
While Stacks provide powerful capabilities, there are some limitations to be aware of during this beta phase:
- Each Stack currently supports a maximum of 20 deployments, which may limit scalability for large environments.
- Stacks have a limit of 500 managed resources per HCP Terraform organization while in public beta. If your organization exceeds this limit, you can no longer apply new resources until you reduce your resource count. If you need more resources, fill out our form to request a resource extension.
- Stacks are not available for users on legacy HCP Terraform team plans. Learn more about migrating to a current HCP Terraform plan.