Managed control plane · deliberate choices

Own the workload.
Trust the boundary.

Learn Amazon EKS from the first Pod to a secure, multi-AZ, observable, cost-aware enterprise platform.

Auto Mode · managed nodes · Fargate · Karpenter · Pod Identity · VPC CNI · delivery · SRE · recovery

REFERENCE TOPOLOGY · THREE AVAILABILITY ZONES
Amazon EKS control and data planes An AWS-managed control plane reconciles workloads running across three Availability Zones with a load balancer and managed data services. AWS-MANAGED EKS CONTROL PLANE AZ · A WORKER WORKER AZ · B WORKER WORKER AZ · C WORKER WORKER ALB · SERVICE · HEALTHY PODSONE APPLICATION SURFACE DESIRE → SCHEDULE → OBSERVE → RECONCILE

What you will be able to do

01 Choose the compute boundary02 Secure people and Pods03 Survive upgrades and failure04 Defend performance and cost

18 production chapters

From Kubernetes objects to platform judgment.

Every stage translates an AWS feature into a decision you can explain, test, and operate.
  1. 01UnderstandRole, Kubernetes, EKS, and compute models4 chapters
  2. 02DesignIdentity, VPC, workloads, scaling, and data5 chapters
  3. 03OperateSecurity, traffic, delivery, SRE, upgrades, and cost7 chapters
  4. 04ProveConnect EKS to SAP, Salesforce, and Workday2 chapters

The first architecture fork

Choose responsibility before instance type.

Auto Mode removes most node work. Managed node groups retain EC2 control. Fargate removes nodes for compatible Pods. Karpenter makes capacity follow workload constraints.

Make the compute decision →
Less node responsibilityAUTO MODEMANAGED NODESSELF-MANAGEDMore node control

Retrieve and apply

Practice consequences, not product names.

A managed control plane is the beginning.

Schedule it. Secure it. Observe it. Recover it.

Start chapter 00