Dojo Terraform

In this dojo, you will learn Terraform basics, how to use the language and how to organize your codebase to make it easly understandable and maintanable.

All along the exercies, we will deploy a basic Frontend / Backend Website exposed on internet with every required AWS infrastructure configurations to make it work.

In order to make it simplier for you, we already set up some parts of the infrastructure and you have a dedicated Virtual Machine with all the tools and rights you need.

Don't hesitate to:

Read every technical comments
Inspect provided documentations
Ask questions

Target Infrastructure

Usefull documentation

This documentation could be useful for the dojo. Don't be afraid to ask questions and search on internet too.

Terraform documentation
Tfswitch
AWS Terraform provider
Visual Studio Code
Slides

Step 0 - I set up my toolbox

We will configure your computer to easly connect to the remote Virtual Machine deployed for you, through Remote-SSH plugin in VSCode.

Technical Comments

Github configuration:

You need a GitHub account and an ssh client
Get your github handle (username)
Make sure you declared your ssh rsa public key in your github account

If not:

Run the following commands (on Linux)

# Create a key pair
ssh-keygen -t rsa -d 4096 -C "your.name@email.com"

# show your public key
cat ~/.ssh/id_rsa.pub

# Copy the value

Follow the doc to add your public key to github

IDE configuration:

Download & install Visual Studio Code
Install Remote - SSH extension for VSCode
Use plugin to connect to VM through <github_handle>@<github_handle>.dojo.padok.school

Remote VM configuration (from VSCode remote-ssh):

Run The following commands:

# Go to ~/dojo-terraform
cd ~/dojo-terraform

# Install required terraform version
tfswitch

# Reload .profile configuration
source ~/.profile

# Check terraform is installed
terraform --version

Access to your AWS Console:

Go to AWS Management console
Click on Sign in
Connect as IAM User
In ~/data.txt find
- aws_account_id
- iam_user_name
- iam_user_password
You are reader on the Account! Feel free to explore!

Acceptance Criterias

I have Visual Studio Code installed with Remote - SSH extension installed
I can connect to my remote VM at <github_handle>@<github_handle>.dojo.padok.school
I can open the repository dojo-terraform on VS Code
I can show my installed terraform version
I can connect to my AWS Console

Know that you are all set up,you are ready to deploy your application with Terraform on an AWS infrastructure!

What did you learn?

Use tfswitch to change your terraform version
Terraform CLI can be used in your console

Step 1 - I Have DNS records for my application

In this exercise, you will deploy a DNS record to make a public endpoint for your future web application. We will have two endpoints:

<github_handle>-frontend.dojo.padok.school
<github_handle>-backend.dojo.padok.school

In the meantime you will learn how to make a basic configuration for terraform.

A - I can deploy a terraform resource

Technical Comments

Provder configuration:

In app/, create a _settings.tf file
Based on AWS provider documentation, configure terraform and provider blocks with
- >= 4.0.0 required provider version
- eu-west-3 region
- default_tags
  - ManagedByTF = "true"
  - User = "<github_handle>"

DNS:

Create a dns.tf file and use aws_route53_record resource
Some useful informations:
- name - value is like <github_handle>-app.dojo.padok.school (for now we juste have one app instead of two frontend and backend)
- zone_id - value can be found in ~/data.txt
- records - value is a list you want to put the lb_dns_name that is in ~/data.txt
- type - value is CNAME

Deployment:

Using your terraform CLI in VSCode console and based on Terraform documentation
Init your terraform configuration using terraform init
You should see on you terminal "Terraform has been successfully innitialized!"
Deploy your configuration
See your DNS record in AWS Console in Route 53 > Hosted zones - dojo.padok.school
Take a look at the .tfstate file created on your app folder

Hint n°1

Have you seen the `Use Provider` toggle on the documentation?

Hint n°2

Have you take a look at the aws_route53_record resource documentation? It seems like some parameters are required...

Hint n°3

Maybe try a `terraform init`, `terraform plan`, `terraform apply` commands?

Acceptance Criterias

I have a DNS record <github_handle>-app.dojo.padok.school in my DNS Hosted Zone
I can make a terraform apply on my app folder
I have a local terraform state with all my deployed resources described in it

At this point, your terraform configuration is set and you can deploy new resources and change existing ones. But some informations are hardcoded in your code, what do you think of that?

B - I keep my codebase DRY

Technical Comments

Data:

Based on AWS provider documentation, use a aws_route53_zone data to refer an existing hosted zone named dojo.padok.school. The route is not private.
Do the same for an aws_lb resource named padok-dojo-lb to get the dns_name of the load balancer.
Refer those data in your record resource. The attributes zone_id and records of your route 53 record should not be hardcoded anymore, but should use data references.

Hint n°1

Be careful! `resources` and `data` are not the same but they are refered with the same name!

Hint n°2

Resources & Data blocks expose some attributes you can use.

Locals:

Create a new file locals.tf and check how locals are declared in terraform
Add a new local applications that is a map of empty object named "frontend" and "backend"
Iterate on this map to create two distinct records
- <github_handle>-frontend.dojo.padok.school
- <github_handle>-backend.dojo.padok.school
Re-deploy your infrastructure. Your Terraform plan will destroy some resources, that's not a problem.
Run terraform state list command to list every resources deployed in your state, and observe how your iterated resources are organized

Hint n°3

Have you look at [terraform types](https://developer.hashicorp.com/terraform/language/expressions/types)?

Hint n°4

To iterate over maps, you can use the [for_each](https://developer.hashicorp.com/terraform/language/meta-arguments/for_each) expression.

Acceptance Criterias

My terraform code does not have external resources hardecoded configuration
I have two DNS records for
- <github_handle>-frontend.dojo.padok.school
- <github_handle>-backend.dojo.padok.school and the code for these resources is unique

At this point, you deployed two dns records for your application, each one pointing on the same Elastic Load Balancer instance. Now let's deploy our web application! :D

What did you learn?

Configure a terraform provider
Deploy a resource
Destroy a resource
Use data to get remote resoures informations
Use locals to avoid repetition in code
Iterate on your resources
List resources present in terraform state

Step 2 - I deploy my application

A - My load balancer can target my instances

We want to configure our Load Balancer to target our application instances, based on their endpoint.

Technical Comments

LB target groups

Based on the official documentation and the following template
Set the accurate configuration to deploy two aws_lb_target_group using iteration
Frontend application runs on port 80 and backend on port 3000
Useful informations are in ~/data.txt
Deploy your resources

resource "aws_lb_target_group" "this" {
  # TODO: iterate to create 2 target groups
  name        = "<github-handle>-<app>" #TOFILL
  port        = ""                      #TOFILL
  protocol    = "HTTP"
  target_type = "ip"
  vpc_id      = "" #TOFILL

  health_check {
    enabled             = true
    interval            = 30
    timeout             = 5
    healthy_threshold   = 5
    unhealthy_threshold = 2
    matcher             = "200"
    path                = "/"
    port                = "traffic-port"
    protocol            = "HTTP"
  }
}

Hint n°1

You may re-use the application local already declared.

LB listener rules:

Based on the official documentation and the following template
Set the accurate configuration to deploy two aws_lb_listener_rule using iteration
Useful informations are in ~/data.txt
Deploy your resources

resource "aws_lb_listener_rule" "this" {
  # TODO: iterate to create 2 listener rules
  listener_arn = "" #TOFILL
  action {
    type             = "forward"
    target_group_arn = "" #TOFILL
  }
  condition {
    host_header {
      values = [<my-endpoint-value>] #TOFILL
    }
  }

  lifecycle {
    ignore_changes = [
      priority
    ]
  }
}

Hint n°1

You may re-use the application local already declared.

Hint n°2

You want to refer the previously created target_group in your resource. A `terraform state list` may help you...

Hint n°2

Feel free to update your existing codebase to avoid code repetition

Acceptance Criterias

I have a target group for my backend and my frontend
My load balancer has new listener rules pointing on the right target groups

B - I use a module to deploy resources

Technical Comments

Use official documentation to get familiar with the usage of a module
Take a look at the service module in modules/service
Some Environment variables must be set
- frontend: BACKEND_URL
- backend: APPLICATION_USER - (that's you!)
Instanciate the module by creating a terraform file in app/. The source can be a relative path to your module.
You need to deploy two services: <github-handle>-frontend and <github-handle>-backend
The ECS cluster name is padok-dojo
Set the proper variables for your module
Useful informations are in ~/data.txt
Deploy your resources

Hint n°1

You may use a data to get ECS cluster details. Its name is padok-dojo.

Hint n°2

We want to enable lb variable and to fill it...

Hint n°3

You might need some outputs from previously created resources to use for the module variables.

Hint n°4

You can re-use the locals created previously to iterate over the module and create the two services.

Acceptance Criterias

My backend endpoint shows the expected data
My frontend endpoints shows a terraform image and congratulates me for my great work 🥳

C - I have my own module for my application

You built a layer for your application in app/ folder. But how would you duplicate your deployment for multiple applications? Let's create a module!

Technical Comments:

First of all, destroy your infrastructure with terraform destroy
Based on official documentation, create a module in modules/app folder
We want to create a module able to deploy one container on one endpoint with its specific configurations.
Re-use the resources you previously deployed
Set your variables and outputs
Use your module in app folder to deploy your frontend and your backend

Hint n°1

You may take as example the service module.

Hint n°2

Avoid declaring data directly in the module. Prefer using variables and declare datas where you call your module.

Hint n°3

Don't forget to keep your code DRY!

Acceptance Criterias:

I have an app module
I use my app module in my layer
I can easly scale up my number of applications instances with multiple configurations

Congratulations!! 🥳 You are a Terraform Builder!

What did you learn?

Use complexe resources
Use resource outputs
Use an existing module
Destroy your terraform infrastructure
Create a module
Create variables

Step 3 - To go further

A - Terraform CLI

Technical Comments:

Based on Terraform CLI documentation and using Treraform CLI
List all resources of your state
Remove a resource from your state and make a plan to see what it will do. You can try to apply to see what happens
Import the resource you deleted in your state
Destroy only one resource with a --target

Acceptance Criterias:

I know some deeper terraform CLI

B - Limitations

Technical Comments:

Take a look at remote state documentation
Take a look at terragrunt documentation
Take a look at Padok team terraform guidelines

Acceptance Criterias:

I know why remote state and state lock are necessary
I understand the limitations of Terraform and know the usefulness of Terragrunt
I have read Padok terraform guidelines

Clean

Don't forget to clean your toolbox by running terraform destroy in exercices/app. It may take a while...

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

Dojo Terraform

Usefull documentation

Step 0 - I set up my toolbox

Step 1 - I Have DNS records for my application

A - I can deploy a terraform resource

B - I keep my codebase DRY

Step 2 - I deploy my application

A - My load balancer can target my instances

B - I use a module to deploy resources

C - I have my own module for my application

Step 3 - To go further

A - Terraform CLI

B - Limitations

Clean

Files

README.md

Latest commit

History

README.md

File metadata and controls

Dojo Terraform

Usefull documentation

Step 0 - I set up my toolbox

Step 1 - I Have DNS records for my application

A - I can deploy a terraform resource

B - I keep my codebase DRY

Step 2 - I deploy my application

A - My load balancer can target my instances

B - I use a module to deploy resources

C - I have my own module for my application

Step 3 - To go further

A - Terraform CLI

B - Limitations

Clean