EKS News 042

Welcome 2023! We hope you all had a restful and happy new year. The start of another year is always exciting for us because it represents a new beginning. We can think of no better way to do that then to start than with the announcement about PrivateLink support for the EKS API, which was first requested in May 2019. Thanks for subscribing to the EKS Newsletter. Your readership and suggestions are important to us.

  • Blue/Green or Canary Amazon EKS clusters migration for stateless ArgoCD workloads
    • This post focuses on an implementation that could allow your platform team to automate the workloads migration from a Blue Amazon EKS cluster to a Green EKS cluster using ArgoCD and GitOps. The solution is comprised of a core-infra stack that creates a shared VPC and shared Route 53-hosted zone for the two blue and green clusters, that are using a local eks_cluster terraform module.
  • Amazon Joins the Open Invention Network
    • Amazon has joined the Open Invention Network (OIN), a patent non-aggression community formed to safeguard essential open source technologies such as Linux and adjacent technologies.
    • Amazon is committing its entire patent portfolio to the body of patents that are free to use with OIN’s defined open source projects. Once the OIN agrees to protect a piece of software, all members are granted royalty-free, community patent licenses for the use of that software. By adding AWS’s patents to the pool, it is helping to reduce the risk of patent aggression for companies that innovate with open source.
    • AWS’s open source investments include hiring dedicated developers and maintainers to work on upstream projects and fine tuning performance of open source projects [like Kubernetes] to run in the cloud.
  • Run fault tolerant and cost-optimized Spark clusters using Amazon EMR on EKS and Spot instances
    • By moving or reusing intermediate shuffle files, you can improve the stability of the Spark [executors] job.
    • This post examines the impact of Node Decommissioning and Persistent Volume Claim (PVC) reuse have on the fault tolerance of Spark [executors] jobs when using Spot.