Creating an EKS Cluster: Version 1.33 from Scratch.

Introduction :

• In-Place Resource Resize for Pods
• Dynamic Resource Allocation (DRA) Enhancements
• Asynchronous Preemption
• Fine-Grained SupplementalGroups Control (Beta)
• ClusterTrustBundles
• Support for Mounting Images as Volumes (Beta)
• Pod procMount Option (Beta)
• CPUManager Policy to Distribute CPUs Across NUMA Nodes
• Zero-Second Sleeps for Container PreStop Hooks (Beta)
• Internal Tooling for Declarative Validation of Kubernetes-Native Types

1. Amazon Linux 2023 (AL2023) & Bottlerocket Only:
   • EKS 1.33 will drop support for Amazon Linux 2 AMIs.
   • AL2023 offers better performance, security, and lifecycle management.
   • Bottlerocket, a container-optimized OS, improves patching, boot time, and security hardening.
2. Improved Custom Resource Definitions (CRDs):
   • Enhanced field selectors for custom resources, making querying more efficient and developer-friendly.
3. Scalability Improvements:
   • Upstream Kubernetes 1.33 is expected to include improvements in scheduler performance and API server throughput, benefiting high-scale EKS clusters.
4. Controller Improvements:
   • Better PVC lifecycle management, helping with StatefulSets and storage reclaiming.
   • More intelligent replica scheduling and taint handling (subject to confirmation in upstream 1.33 notes).
5. Security Enhancements:
   • Continued deprecation of older APIs and increased emphasis on RBAC hardening.
   • Expect new features around network policy support and node isolation.

Disadvantages of EKS 1.33

1. No More Amazon Linux 2:
   • Users relying on AL2 will need to migrate to AL2023 or Bottlerocket, which may involve effort in testing and compatibility validation.
2. Potential Breaking Changes:
   • New versions often deprecate beta APIs or change CRD schemas, requiring app manifests to be updated.
   • Could require helm chart or controller updates.
3. Upgrade Complexity:
   • Upgrading from 1.31 or earlier to 1.33 might involve multi-step upgrades (1.31 → 1.32 → 1.33), increasing downtime and testing scope.
4. Early Bugs/Instability:
   • The first few weeks of a new EKS version rollout may expose edge case bugs, especially for large or complex clusters.
5. Learning Curve with New AMIs:
   • Bottlerocket and AL2023 require different tooling for customization, logging, and monitoring compared to AL2

Step 1: Login AWS console

• Open IAM Dashboard
• **** Create a user. username : ashish
• Attach AdministratorAccess policy.
• Create access and secret key.

Step 2: Create a Ec2 machine

1. Open a EC2 Dashboard.
2. Launch instance
3. Name and Tags : MyTest
4. Application and OS Image ( AMI ) : Amazon Linux 2023 AMI
5. Instance Type: t2.micro
6. Keypair : ashish.pem
7. Network Settings : VPC, subnet
   • Security Group : 22 - SSH (inbound)
8. Storage : Min 8 GiB , GP3
9. Click Launch instance

Step 3: Login EC2 instance and configure Access/Secret key.

Step 4: Install eksctl and kubectl using AWS CLI.

• Download and extract the latest release
• Move the extracted binary to /usr/local/bin
• Test that your eksclt installation was successful.

• Download kubectl version
• Grant execution permissions to kubectl executable
• Move kubectl onto /usr/local/bin
• Test that your kubectl installation was successful

Step 5: Create a cluster using eksctl command.

• eksctl create cluster : Creating a cluster​​ eksctl
• --name ashish :**** Name of Cluster
• --version 1.35 : EKS cluster version
• --region us-east-1 : AWS Region Name
• --nodegroup-name ashish-workers : Autoscaling Group Name
• --node-type t3.medium : instance type
• --nodes 2 : Desire Node capacity is 2.
• --nodes-min 1 : Minimum Node capacity is 1.
• --nodes-max 3 --managed : Maximum capacity is 3.

1. AWS CLI
   • Check how many pods are running

1. AWS Console

• Verify EKS Cluster and version.