Skip to main content

Deploy Everything

After you've written your components and configured your stacks, now we're ready to deploy them!

You will learn

  • How Atmos identifies components using context variables and naming patterns
  • How to preview your configurations and changes before applying them
  • How to deploy every Terraform component in a selected stack

Deploy the Stack​

Before deploying, validate the stack manifests:

atmos validate stacks

You can also orient yourself with the built-in list commands:

atmos list stacks
atmos list instances -s dev

Preview every Terraform component in the dev stack:

atmos terraform plan --all -s dev

Deploy every Terraform component in the dev stack:

atmos terraform deploy --all -s dev

When you're done, you can destroy every Terraform component in the stack:

atmos terraform destroy --all -s dev -auto-approve
Let Atmos prompt you

For single-component Terraform commands, you can name the component explicitly or omit the component and stack in an interactive terminal so Atmos can prompt you:

# Explicit
atmos terraform plan station -s dev

# Interactive: choose component, then stack
atmos terraform plan

For bulk operations, use explicit selectors such as --all, --components, or --query.

Stack Search Algorithm​

Looking at the commands above, you might have a question "How does Atmos find the component in the stack and all the variables?"

Let's consider what Atmos does when executing the command atmos terraform deploy --all -s dev:

  • Atmos uses the stacks.name_template defined in the CLI config. In this example, we have defined a simple name based on stacks.name_template: "{{ .vars.stage }}". This means that the stack name is just the stage part of the stack name.

  • Atmos searches for the stack configuration file (in the stacks folder and all sub-folders) where stage: dev is defined (inline or via imports). During the search, Atmos processes all parent (top-level) config files and compares the context variables specified in the command (-s flag) with the context variables defined in the stack configurations, finally finding the matching stack

  • Atmos finds every Terraform component in the stack. In this simple Quick Start, that means Atmos resolves the station component by processing all the inline configs and all the imports

  • Atmos deep-merges all the catalog imports for the station component and then deep-merges all the variables for the component defined in all sections (global vars, terraform vars, base components vars, component vars), producing the final variables for the station component in the dev stack

  • And lastly, Atmos writes the final deep-merged variables into a .tfvar.json file in the component directory and then executes the Terraform command

Ready to take the next step?

Now that you’ve seen Atmos in action, you can explore the extra credit!

Or take a moment to explore its core concepts. You have only just scratched the surface of Atmos. Atmos is a powerful enterprise-grade framework with so much more to offer!