cloud-gitops-examples

This repo serves two purposes:

• gitops tooling examples with terraform, argocd, etc.
• demonstrate various cloud topologies with gloo products

Available Modules

module	argo apps	description
vpc-one-region		1 region vpc
vpc-two-region		2 vpcs with peering
redis-1region	values-aws-core-infra.yaml values-redis-tester.yaml	1 region vpc, elasticache single-node cluster with auth and tls, single-node eks cluster
redis-2region	values-aws-core-infra.yaml values-redis-tester.yaml	2 peered vpcs (2 region), elasticache in 2-region global mode with tls and auth, single-node eks cluster in each region
eks-1region	values-aws-core-infra.yaml	1 region vpc, single-node eks cluster

Requirements

• terraform (I suggest https://tfswitch.warrensbox.com/ to manage versions)
• an AWS account with a working CLI env (can you aws sts get-caller-identity?)

Environment for terraform state storage

State is stored in s3 (use a private encrypted bucket).

export AWS_TF_BUCKET="your-aws-bucket-for-terraform-state"
export AWS_TF_REGION="your-aws-region-for-terraform-state"
export TF_STATE_KEY_PREFIX="your-team-name"

What is STACK_VERSION.txt?

It is a best practice to name resources in terraform with prefixes that allow multiple deployments of the same module. Think of a collection of terraform modules (vpc, eks, redis, etc.) as an "infrastructure stack" that is defined with infrastructure as code (IaC). Because it is "just code", it can be versioned. Because it's versioned, multiple versions can be deployed in parallel.

The STACK_VERSION.txt file contains a string prefix that will be used on all aws resources. This allows you to experiment with changes to infrastructure via terraform without breaking existing deployments, or simply to deploy the same module multiple times on the same account. Use of a resource name prefix is NOT a terraform feature; care is taken by module developers to use input variables like stack version to name resources accordingly.

Module-specific environment requirements

Some modules require additional env vars (like redis_auth in the redis modules). View the README.md in the respective module folder, or simply run ./terraform-wrapper.sh init && ./terraform-wrapper.sh plan in the module directory to learn what required vars are missing. Any TF var can be automatically injected by setting env vars with the prefix TF_VAR_ - for example TV_VAR_redis_auth would set the value for the redis_auth var.

How to install a module

1. Validate your aws cli env with aws sts get-caller-identity
2. Change directory to the module you want to install cd terraform/vpc-one-region
3. Set bucket, region, and key prefix values:

   export AWS_TF_BUCKET="your-aws-bucket-for-terraform-state"
   export AWS_TF_REGION="your-aws-region-for-terraform-state"
   export TF_STATE_KEY_PREFIX="your-team-name"

4. Use ./terraform-wrapper.sh to interact with terraform:

   ./terraform-wrapper.sh init
   ./terraform-wrapper.sh plan

5. Validate the terraform plan to be executed
6. Apply with ./terraform-wrapper.sh apply

Creating new modules

See ./CONTRIBUTING.md