Terraform: what is it and why should you care?

Terraform is an infrastructure configuration tool created by HashiCorp that allows you to create, change, and improve your infrastructure in an easy-to-use and consistent manner. But, what exactly does this mean? How does it work? What does it actually do? In this guide, we’ll give you all the details on how terraform works and how you can use it to manage your infrastructure with ease!

Introduction to Terraform

When you think of designing data centers, chances are, infrastructure as code comes to mind. Terraform allows developers and administrators to create data center resources using a simple programming language that can then be stored in version control like any other code. That sounds great—but wait, there’s more! Terraform also simplifies ongoing management of your infrastructure—as changes occur over time—by treating your existing cloud resources as immutable golden images for each provider. This means that when something changes in AWS or Azure or Google Cloud Platform, your Terraform configuration will tell your cloud provider how to change things with little-to-no manual intervention from a human engineer. Most importantly, though, it gives you all of these features at an API level so that even if someone else has already implemented all of these features into their application or another configuration tool, anything built on top of Terraform works seamlessly with any provider (like Ansible), deployment platform (like Packer) or cloud software (like Jenkins). Want to use popular developer tools with built-in support for providers? Go ahead! Need some extra fine tuning from a proven system administrator rather than have a developer write those scripts themselves? Great idea; let them focus on building stuff and let someone else deal with infrastructural maintenance.

Introduction to infrastructure as code

Infrastructure as code refers to building, configuring, and maintaining your infrastructure using code rather than manually. To manage your cloud deployments, AWS provides its own application called Terraform (which we’ll talk about more later). The easiest way to get started with Terraform is to use a pre-existing template. This will help you avoid common mistakes in configuration. You can then tweak these configurations to meet your needs once you’re up and running. A good example of such templates is found here . While building from scratch may be easier for some people, trying to troubleshoot problems without templates increases complexity. Templates are also an easy way to share your work with others who are interested but not yet familiar with your environment. At my company, for example, we routinely share our Terraform files so that engineers who aren’t directly involved in creating or modifying them know what’s going on and can provide input when necessary (and sometimes when not necessary!). And lastly, templates are designed by people who have already done all of their own learning—thereby avoiding having to do all of that yourself!

How do we use terraform?

Using terraform allows us to declaratively build, change, and version infrastructure. This post is going to go over how we can use terraform with Ansible to manage our cloud infrastructure. Terraform will let us define an Ansible inventory file that will contain all of our hosts we need to deploy on. We’ll then use Terraform to turn that into real hosts on AWS (or any cloud provider) using a simple command like terraform apply . If we ever need to redeploy or make changes down the road, we can just run terraform apply again. Our host configuration management tooling stays in sync with our actual configuration! You get to have your cake and eat it too. What are some best practices for working with terraform?: When I first started working with Terraform there were a few things I wish I knew about early on. Understanding these best practices can help improve workflow and debugging in more complex environments: Even if you have dynamic IP addresses from AWS, at least lock them in to prevent other engineers from accidentally removing them. Set up health checks so that when something goes wrong or gets lost it’s easier to detect. Store each piece of infrastructure in its own file but keep those files organized sensibly so they’re easy to read through later.

How does Terraform work

Terraform orchestrates building, changing, and versioning infrastructure safely and efficiently. It reads configuration files that describe how all of your systems should be set up, in a declarative language. Then it brings them up or down in coordinated changes based on those configurations. The power here is in Terraform’s flexibility to integrate with your existing tooling—from simple shell scripts to advanced development environments like Chef or Puppet—to implement operational processes that fit into your workflows already. Because Terraform runs alongside other tools, you can use it to do a little bit of infrastructure as code (i.e., make changes automatically), while still leveraging third-party libraries for more complex tasks. This gives you an infrastructure light approach to control your stack, by combining your chosen tools in flexible ways. This approach also gives you powerful visibility into what Terraform does for you versus what its supporting components do. For example, if an upstream package update causes problems, it will be clear which component is at fault (Terraform code vs. underlying software). In addition to these benefits, using Terraform helps reduce lock-in since there are many supported providers available from different vendors. Further, because Terraform makes managing services explicit in its configuration files rather than relying on manual steps by operators during upgrades and maintenance windows, resource utilization will likely increase without adding any additional staff.

How does it differ from AWS CloudFormation

Terraform has many similarities to AWS CloudFormation (and to Microsoft Azure Resource Manager), which may lead one to wonder how they differ. Terraform uses a provider-driven approach, where each resource needs a plugin in order for Terraform to manage it. So, for example, if you want to create an Amazon Web Services (AWS) EC2 instance with Terraform, you’ll need a provider plugin that provides EC2 as a service within your infrastructure. In contrast, CloudFormation can create resources directly without any such plugins. This means that you could use CloudFormation to create an EC2 instance—or even install some software on it—with no additional work on your part. You might not even know that there are separate resources involved behind the scenes; these details are abstracted away from you, just like with Terraform. However, using the service-provider model means you have full control over all aspects of your infrastructure stack. This can be good or bad depending on your situation; CloudFormation is certainly easier to get started with but lacks flexibility down the road when things start getting complex enough to require custom code and logic outside of Terraform’s capability set. It’s worth noting that CloudFormation uses JSON templates while Terraform utilizes HCL. Because Terraform compiles to native machine language before creating your infrastructure, its performance is superior compared to CloudFormation. And unlike CloudFormation templates, which only run once per environment configuration change during deployment, Terraform modules can run multiple times at different points in time during application lifecycle management. The difference comes into play with stateless applications and infrastructure changes made due to API versioning requirements or other non-versioned changes (which also include CDN modifications). Essentially every run of a terraform plan command results in actual changes being made to your infrastructure.

How can we help our teams use terraform effectively

First, let’s break down what is terraform into its two core components—Terraform is a tool for automating your infrastructure. Secondly, with Terraform you can plan or configure resources to create infrastructure that closely matches your plan. This post will discuss how automation can help teams work faster and more efficiently by freeing them from repetitive tasks like provisioning servers, deploying applications, or rebuilding production environments after a crash. To accomplish these goals we need to take a step back to understand where Terraform fits in within an enterprise architecture . Once we have done that then we can determine how to use it effectively. With services such as Amazon EC2, Google Cloud Platform and Microsoft Azure making up a large portion of most enterprises architectures today there are some questions that need to be answered before using Terraform effectively—do I provision AWS instances with Ansible or Terraform? How do I update existing infrastructure using Terraform when my team doesn’t know how to program yet? Should I use cloud-init in conjunction with Terraform if so which do I execute first and last ? As you can see there are multiple levels of configuration management at play here all of which must be working together smoothly if any one system alone is going to get used properly.