🐾Best practices for storing application code and Terraform code🐾

❓How should you store your application code for Lambda functions or applications running on ECS? Should you use the same or separate repository from your Terraform configuration?

✅ There are no official best practices for that, but I can share my experience and thoughts with you. In general, I prefer to store both application and Terraform code in one repository because:

1️⃣ I consider Terraform configuration and application code as one deployable unit, which should be stored together. Also, it is much easier to deploy application into the cloud with CI/CD tools, when it is one repository.

2️⃣ It is more convenient to have changes in configuration and application related to the same reason in one commit. I saw pretty similar opinion from HashiCorp guy in Discuss forum: “If two things will often change together then it’s helpful for them to be represented together in a single repository.”

3️⃣ I think all IaC tools have the same goal and therefore can share best practices. AWS CDK has its own best practices and judging based on what I have seen in example repositories, usually, application code is stored in the same repository as configuration code.

So, as most of the time I use Terraform + Terragrunt configuration for resources and deployments, usually my repository has the following structure:

Workloads/ 
|-- vpc/ 
|   |-- main.tf 
|   |-- README.md 
|-- some_workload/ 
|   |-- main.tf 
|   |-- assets/ 
|   |   |-- application_code.py
|   |-- README.md 
Deployments/ 
|-- prod/
|   |-- us-east-1/ 
|   |   |-- vpc/ 
|   |   |   |-- terragrunt.hcl
|   |   |-- some_workload/ 
|   |   |   |-- terragrunt.hcl
|   |-- eu-central-1/ 
|   |   |-- vpc/ 
|-- dev/
|   |-- vpc/ 
|   |   |-- terragrunt.hcl

🤓Let me know which approach you use for your IaC repositories in the comments.

---

If you like this post, you can share APAWS newsletter with friends:

---