Infrastructure Engineer Code Challenge

Using Bash, Kubectl and Terraform; create the following configuration.

Please use variables so we can quickly create many different environments.

• GKE Autopilot Cluster
• Nginx deployment
• HPA for Nginx deployment to autoscale on cpu (50%)
• Cloud SQL MySQL 8 Instance
  • Machine Type: standard
• Cloud SQL database
• Seed database with the table db.user
  • id
  • first_name
  • last_name
  • email
• Cloud Memorystore Redis Instance
  • Capacity: 1GB

Prerequisites

CLI Tools:

$ terraform -version
Terraform v1.1.2
$ gcloud version
Google Cloud SDK 330.0.0

Provider Auth

$ gcloud auth application-default login
$ ln -s /home/$USER/.config/gcloud/application_default_credentials.json credentials.json

Possible Improvments

• All the region and project vars for the different GCP resources should have the option to override based on a variable but default to the global project and region vars. For example:

  region = var.redis_instance_region != null ? var.redis_instance_region : var.region
  

• Split the different atomic units/resources into their own modules to make them more usable. For example, a k8s module, a database module, a cache module.
• Use some of the modules provided by Google to better implement best practices as set out by them.
• Add validation to the varaibles to ensure they are in the correct format.
• A remote state to backend.
• Use templates and variables to generate the schema.sql file.
• Improve HA on the database and redis instances.
• Manage the VPCs.