Creates a Kubernetes cluster in Auto mode that uses AWS-managed nodes.
- SaaS workloads
- stateless services
- CI runners
- internal platforms
- preview environments (👀 very relevant to you)
- cost optimization focus
- teams without deep k8s infra skills
- custom networking
- heavy DaemonSet reliance
- GPU tuning
- special storage drivers
- low-level performance control
- platform engineering teams wanting full control
- highly predictable infra topology needs
AWS automatically manages the following addons:
- VPC CNI
- CoreDNS
- kube-proxy
- Pod Identity Agent
- metrics server
- networking dataplane components
In auto mode, AWS manages a fleet of nodes which prevents you from performing various operations:
- You cannot run DaemonSets.
- You cannot SSH into nodes.
- You cannot optimize pod scheduling; AWS optimizes this.