Article originally published on LinkedIn on 2024-10-28.
Source: https://fr.linkedin.com/pulse/terraform-5-technique-pour-am%C3%A9liorer-la-meddy-menzikoff-mtemf

For several years, Terraform has been one of my main tools. An Infra as Code tool capable of supporting different Cloud services, it also offers numerous features making the tool extremely powerful.

In this article, we will cover some of the most essential tips and tricks for writing good, easy-to-use, and scalable Terraform code.

Variables#

You can use variables to configure resources or to maintain environment-specific settings. Variables can also be used to create default configurations.

The best practice is to define the variables in a separate file in a separate file that can be adapted depending on the environment. You can then reference these variables in your resource definition using the ${var.variable_name} syntax. These variables can be passed into modules, allowing you to easily reuse your code with different variations.

Modules#

With modules, you can group related resources and configurations and package them for reuse across projects or even within a single project. They also allow you to create custom abstractions, making your code easier to read and maintain.

Modules are particularly beneficial when you need to deploy similar infrastructure components across multiple environments. Instead of duplicating your code, you can extract it into a module and then use it in different environments by adapting the settings. This method not only reduces duplication, but also simplifies the management of your infrastructure by ensuring consistency between environments.

The best part is that Terraform modules are simple to create and use. Just define your module in a separate directory with specific inputs and outputs. You can then reference it in your configuration file using a module block. A valuable resource for finding ready-to-use Terraform modules is the Terraform Registry, which offers a wide range of modules developed by the Terraform community that you can use and enrich.

Conditionals#

Conditional logic is crucial for designing more sophisticated infrastructure scenarios. Terraform provides various conditional statements to selectively apply resources or generate configurations based on certain conditions. For example, you can use a condition to create a resource only if a specific environment variable is set, or to create a resource only in a given region.

To use conditional statements in Terraform, you can use if or for_each expressions. You can also use the count parameter, which allows you to create multiple instances of a particular resource based on the specified number. Using conditional logic will help you refine your infrastructure configuration and make it more efficient.

Validation and Linting#

Terraform offers validation and linting tools that help you ensure your code is both efficient and error-free. For example, running terraform plan gives you a preview of the changes Terraform plans to make, while terraform validate checks your configuration file for possible syntax errors.

In addition to the built-in tools, you can also combine your Terraform code with linting and validation tools such as Terrascan or tfsec for further verification. These tools are able to detect common problems and errors that might go unnoticed.

Remote state storage#

Terraform uses state files to manage the state of your infrastructure. By default, Terraform stores state files locally. However, this approach can be risky in the event of a crash or loss of your local machine. Additionally, it is difficult to share state files with other team members during a collaborative project.

It is therefore recommended to store your state file in a centralized location, such as an S3 bucket or a remote database. This allows multiple team members to collaborate on the same infrastructure without creating conflicts. To store your state remotely, you can configure Terraform to use a backend, such as Amazon S3, Azure Blob Storage, or Consul from HashiCorp.