Kubernetes on an AWS-hosted cluster (not EKS): Where does infrastructure replication logic go?

3/29/2019

I have a Kubernetes-orchestraded infrastructure that is served on an AWS-hosted cluster. I'd like to have routines that would allow me to spawn similar infrastructures. The difference between the original infrastructure and the newly-spawned ones would mainly be the DNS's used and the images that I would serve.

My question is: What is the most appropriate place for this similar-infrastructure-spawning code to reside: Kubernetes ? My CI/CD tool, Drone? Some other DevOps stack component of which I'm not even aware?

-- bsam
amazon-web-services
drone
kubernetes

1 Answer

3/29/2019

Do you ever think about InfraAsCode tech.

You can directly develop your infrastructure using code like:

  • Cloudformation (AWS)
  • Terraform (multi provider)
  • Ansible
  • ...

You will then be able to configure all providers services (not only your cluster)

You will then be able to deploy it with a single command and use parameters on it.

Otherwise you can also you a tool like Kops (https://github.com/kubernetes/kops) to automate deployment of you K8s cluster.

Once you choose the right tool, you will then be able to source it using Git repository or whatever.

-- Steve HOUEL
Source: StackOverflow