This repo contains my Terraform scripts that deployes all Kubernetes manifests to my k8s clusters. Providers used to deploy manifests to Kubernetes.
- Official Kubernetes provider for k8s resources.
- Alpha Kubernetes provider for generic k8s manifests.
- Auth0 provider to create clients and resource servers.
- Community Helm Provider for deploying Helm charts.
- Official AWS provider to handle Route53 registration of applications.
The terraform scripts does NOT contain the provisioning of Kubernetes itself. I have two clusters running, one in Amazon EKS deployed by cloudformation and one on-prem on VMWare deployed with Tanzu Kubernetes Grid (TKG).
I run several deployments at the same time, both on-prem in parallel and in Amazon EKS.
Under the directory envs there is one directory for each deployment.
The different development environments for software development.
- prod is deployed to Amazon.
- stage is deployed to Amazon.
- test is deployed to on-prem on VMWare Tanzu.
- dev is deployed to on-prem on VMWare Tanzu.
In addition
- services is a shared services on-prem on VMWare Tanzu.
To deploy an environment its meant to stand in its directory and do terraform apply
Secrets are stored in encrypted files in the repo and contains a couple of variables
that are needed to run the terraform scripts.
Two different default templates to configure an environment with.
Is a small and mostly empty base template. It contains common stuff like a default configuration of traefik and forwardauth and not much more.
Is a more specialized template that contains some applications that are meant to be installed just once for each cluster and contains tools for cross-cutting concerns like logging and observability implemented with prometheus, grafana, loki and more.
Under modules I have put all my re-usable code, the environments should not contain much code itself but should instead use the modules to configure deploy the software.