Infrastructure Take Home

Treat this system as a production system.

Getting Started

Clone this repository locally. Create your own public git repository in github or somewhere we can access and push this code into it. Make changes to your repository. Getting things to work for you is part of the assessment.

You will be assessed by someone cloning your repository when you're finished and running your instructions to recreate the expected solution. If we cannot run your repository instructions we cannot assess your work.

Prerequsites

You will need the following:

• docker runtime and tools
• k3d CLI
• opentofu binary or terraform
• kubectl binary
• git

Starting point

Use terraform or opentofu to initialise a k3d cluster and postgres instance locally from the tofu directory. Install Argo CD into the k3d cluster by following the instructions in the argocd directory.

Problem

Please add commits to your fork of the repo to answer this problem. Note: the use of the word postgrest is confusing, but correct - this is a project that we're going to deploy.

Add a user to the database

Please add a super user to the postgrest database.

Inject a secret for postgrest

Creating a superuser account in this new database, inject the secrets into the k3d cluster into a namespace called postgrest. You must do this with terraform/opentofu.

Install Postgrest into the k3d cluster

https://docs.postgrest.org/en/v14/

The result should be an accessible endpoint that you can use in your browser.

Inject some data from the cluster using a Job

Use a kubernetes job to inject some data into the postgres database

Provide an expected screenshot

Update this file, README.md, with a screenshot of what we should see when we visit the URL after following your instructions - this should show us the data you have injected.

Setup

Full instructions are in docs/setup.md. Quick reference in docs/running-short.md.

Two things worth noting:

ArgoCD wait — wait for all deployments, not just argocd-server:

kubectl wait --for=condition=available deployment --all -n argocd --timeout=120s

ArgoCD uses a separate argocd-repo-server to clone git repos and read manifests. If you only wait for argocd-server (which starts faster), the first sync attempt fails because the repo-server isn't ready yet — then ArgoCD waits ~3 minutes before retrying.

OpenTofu — the kubernetes provider was replaced with terraform_data + local-exec kubectl. The kubernetes provider validates the kubeconfig context at init time, before the cluster exists, which causes tofu apply to fail on a fresh machine. Using kubectl inside local-exec avoids this — it only runs during apply, after the cluster is created.

Expected Result

Visit http://localhost:8080/employees after running the setup instructions.