polyqube

A Pulumi solution for deploying hybrid K3S clusters to AWS, Azure, and GCP.

Polyqube fills in the gap for organizations that are unable to use cloud provider specific Kubernetes distributions, like EKS, while also providing similar features found by cloud provider Kubernetes distributions.

Features

High-Availability K3S deployments
GitHub actions for deploying new clusters
Packer and Ansible for creating golden templates for new K3S releases
Configures cloud provider application and network load balancers for your clusters
- Application load balancers for your Kubernetes services and ingresses
- Network load balancers for load balancing traffic to the control plane nodes
Supports hybrid clusters where worker nodes exist outside of the cloud platform
Rolling upgrades of K3S versions through VM Template red/blue/green deployments
Developer and Operator friendly CLI for creating new cluster definitions
Support for ArgoCD for bootstrapping and managing Kubernetes resources on all the clusters

Getting Started

Pulumi CLI
- >= v3.115
Go
- >= v1.22
Packer
- >= v1.10.1
Ansible
- >= v2.16.3
Create a Deploy token on the repository that contains your ArgoCD manifests
- This is required for ArgoCD to have read access to the kubernetes manifests
Cloud provider credentials

Create a K3S VM Template

Here is an example using Packer to create a K3S VM Template for AWS

# Init the project
packer init aws-k3s-control-plane.pkr.hcl

# Validate the project
packer validate aws-k3s-control-plane.pkr.hcl

# Build the AMI
packer build aws-k3s-control-plane.pkr.hcl

Create a Cluster Definition and Stack

cd cloud_resources/

# CLI will create a valid cluster definition and Stack for the cluster
go run cmd/cluster/main.go

Bring up the Cluster

pulumi stack select <cloud platform and region>
pulumi up

dobecad / polyqube