This set of scripts and TF files can deploy and destroy a sample load-balanced Fargate cluster, driven from a small set of input files.

0.0 - Download the files to a local directory, keeping the relative directory structure intact. suggested path: ~/validity-homework

0.1 - Run 00_config_console.sh to install the required components if needed, which are: unzip, terraform, python, aws cli.

0.2 - Edit pilot_secrets.tfvars to contain a valid access key and secret key.

# this is used for backend and provider
access_key="A******************A"
secret_key="*************************************XyZ"

0.3 - Edit pilot_inputs.tfvars to specify the hosted zone id and domain name of a public Route53 zone. You should not need to edit any other inputs.

pilot_inputs = {
  name = "validity-homework"  # name of this project; will be used to create a CNAME for the ALB hostname
  site = "Ireland"            # aws region name for the deployment
  cpu  = "0.25 vCPU"          # container compute allocation
  ram  = "0.5GB"              # container memory allocation
  hzid = "*************"      # hosted zone id for the CNAME
  domain = "***********"      # hosted zone domain name
}

0.4 - Edit the ./tf/backend_s3.tf file to specify a valid S3 bucket and key path for the S3 backend, OR leave the file commented out to use the local /validity-homework/tf directory as the backend.

# terraform {
#   backend "s3" {
#     bucket      = "<a-valid-bucket-name>"
#     key         = "<a-valid-key-path>/terraform.tfstate"
#     region      = "<aws-region-for-the-bucket>"
#   }
# }

1.1 - Execute 01_terraform_plan.sh. It should run successfully, performing a terraform init and terraform plan, and generating the following files, in the same dir as the script file:

• tf.plan
• tf_plan.log

2.1 - Execute 02_terraform_apply.sh. It should run successfully, performing a terraform init and terraform apply, and generating the following files, in the same dir as the script file:

• tf_apply.log
• az_count.log
• try_this_address.log

2.2 - try_this_address.log should contain the URL that matches the values in the pilot inputs, similar to http://validity-homework.infra.ac:80. Clicking or CURL'ing to this URL repeatedly should produce output similar to the following...

`I'm 4171eb990bc8 running on linux/amd64`

... with a finite number (3) of different IDs showing up in the responses, one for each available Availablity Zone in Ireland/eu-west-1 at the time of deployment. Check the output of az_count.log to view the available AZs.

Note: I noticed that clicking the URL in a browser seemed to exhibit what looked like 'round-robin' behavior with the ALB routing, while running a batch of curl commands from bash would exhibit different behavior that look like weighting/random, which is what I expected. Was it a fluke?

2.3 - (optional) Use the AWS CLI or log in to the AWS web console and inspect the deployed resources.

3.1 - (optional) Execute 03_terraform_show.sh. It should run successfully, performing a terraform init and terraform show, and generating tf_show.log in the same dir as the script file.

3.3 - (optional) inspect the terraform show output file.

4.1 - Execute 03_terraform_show.sh. You will be prompted to confirm terraform destroy. It should run successfully, performing a terraform init and terraform destroy, and generating tf_destroy.log in the same dir as the script file.

4.2 - (optional) Use the AWS CLI or log in to the AWS web console and confirm resource destruction.

It's important not to leave sensitive data around:

5.1 - Delete the pilot_secrets.tfvars

5.2 - Delete the tf.plan

5.3 - Delete the generated .log files

5.4 - Delete the contents of the /tf/.terraform/ directory