Health To Go Infra

This repository is capable of deploying a new AWS EKS cluster and deleting it right after using Terraform.

Getting Started & Documentation

When running Terraform locally, it is important to have tfenv. This tool works for terraform like pyenv works for python. It allows you to have multiple different terraform environments using different terraform version. Follow the README from tfenv to install it for your respective OS.

After installing tfenv and having the PATH properly configured, run tfenv install. It looks for the .terraform-version file.

Now you should have terraform available, you can test this by running terraform --version.

Projects and Modules

This repository, for the time beeing, has the projects: eks and kubeconfig. You can identify it by the folders name. For the time beeing, it also has eks and k8s modules within the modules folder.

Never run terraform within the modules folder. It is intended to be run inside each project individually, but not inside modules. The modules will be used by the projects when you run terraform within the projects. Also, run eks project always before kubeconfig, the kubeconfig needs the cluster to be existent.

The Steps

1. First run terraform init, and terraform workspace list inside eks folder. This command will display all the workspaces existent. You must create a new workspace if you wish to create a new cluster. terraform workspace new <WORKSPACE_NAME>. The workspace name will be used by terraform as prefix for the cluster name.

2. After setting the workspace, run terraform plan. This will simulate what terraform would do in a real terraform apply as a safety measurement.

3. Later, run terraform apply (MAKE SURE YOU VALIDATED THE RESULTS OF THIS ACTION WITH terraform plan), it will generate all the resources necessary to run an EKS cluster. When you create an Amazon EKS cluster, the IAM entity user or role, such as a federated user that creates the cluster, is automatically granted system:masters permissions in the cluster's RBAC configuration in the control plane. This IAM entity doesn't appear in the ConfigMap, or any other visible configuration, so make sure to keep track of which IAM entity originally created the cluster.

You will need the endpoint, certificate, and cluster name generated by EKS in order to use kubectl (This is not necessary now). You can extract this information in the AWS Console by accessing your recently created EKS or you can also access this data programatically using the EKS module output from this repository.

The kubeconfig yaml follows. Please replace ENDPOINT, CERTIFICATE-AUTHORITY-DATA, and CLUSTER-NAME

apiVersion: v1
clusters:
- cluster:
    server: <ENDPOINT>
    certificate-authority-data: <CERTIFICATE-AUTHORITY-DATA>
  name: kubernetes
contexts:
- context:
    cluster: kubernetes
    user: aws
  name: aws
current-context: aws
kind: Config
preferences: {}
users:
- name: aws
  user:
    exec:
      apiVersion: client.authentication.k8s.io/v1alpha1
      command: aws-iam-authenticator
      args:
        - "token"
        - "-i"
        - "<CLUSTER-NAME>"

4. On the kubeconfig folder run terraform init and terraform workspace list. You must also create a workspace for it and use the SAME EXACT NAME you have given the eks project.

5. You will need to run terraform import module.k8s.kubernetes_config_map.aws-auth kube-system/aws-auth to import the aws-auth into your terraform state.

6. Finally run terrafom plan and terraform apply (PLEASE SEE THE WARNINGS AND RECOMMENDATIONS ABOVE IF YOU HAVEN'T ALREADY).