This is our last edition before the upcoming holidays. We’ll return again in 2023. On behalf of all of us here at AWS, enjoy the holiday season and have a merry new year. For fun, we’ve included a couple links to stories about ChatGPT in this edition, the conversational AI that is both amazing and a little disconcerning. There are also lots of interesting GitHub projects to feast on in this edition, including the ASCII visualization tool, eks-node-viewer, that was used to demonstrate the effectiveness of Karpenter consolidation. Finally, EKS recently added support for passing custom configuration to addons at cluster provisioning time which aims to accelerate the creation of “application-ready” clusters.
- The EKS Newsletter Editorial team
- Amazon EKS launches automated provisioning and lifecycle management for Windows containers
- Managed node groups now supports Windows worker nodes. To learn more about EKS Windows managed node groups and how to get started, visit the EKS public documentation .
- Amazon EKS add-ons now supports advanced configuration
- EKS now supports advanced configuration of cluster add-ons, e.g. Amazon VPC CNI, CoreDNS, etc, enabling you to customize add-on properties to help you meet performance, compliance, or additional requirements not handled by default settings. For more information on passing advanced configuration at provisioning time or after the cluster is created, please see the accompanying blog post on Amazon EKS add-ons: Advanced configuration .
- Amazon EMR on EKS now supports Nvidia RAPIDS Accelerator for Apache Spark
- Starting with EMR 6.9, EMR on EKS is introducing a new Nvidia RAPIDS Accelerator for Spark image. Customers can use the same StartJobRun API to run their Spark jobs, and simply specify a new Spark-RAPIDS release label to leverage RAPIDS Accelerator on an EKS cluster with GPU supported instance type. To learn more about how to run Nvidia RAPIDS Accelerator for Apache Spark, please visit the documentation page.
AWS container blogs
- Deploy geo-distributed Amazon EKS clusters on AWS Wavelength
- EKS lets you deploy workloads across all the Availability Zones (AZs), Local Zones, and AWS Wavelength Zones within a given region. Local Zones and AWS Wavelength Zones offer compute and storage services with lower latency than traditional AZs can provide.
- AWS Wavelength Zones are logically isolated data centers, within the telecommunication providers’ networks, that are connected back to the AWS Region via redundant, low latency, and high-throughput connectivity. They are accessible only via the wireless network of the partner supporting the AWS Wavelength Zone.
- This post describes how Amazon EKS can be used to deliver highly geo-distributed applications for low-latency application access using Wavelength Zones. It captures 2-years-worth of customer feedback, best practices, and holistic lessons learned from their environments.
- Expose Amazon EKS pods through cross-account load balancer
- This blog explains how to expose Amazon EKS pods using a “cross-account load balancer” in which a load balancer in one AWS account routes traffic to pods in another AWS account using a shared VPC network architecture.
- Blue/Green Kubernetes upgrades for Amazon EKS Anywhere using Flux
- Amazon EKS-A cluster upgrades are performed in place using a rolling process and upgrades can only happen one minor version at a time (e.g., version 1.20 to 1.21). The blue/green deployment method allows you to shift traffic between two nearly identical environments that are running different versions of Kubernetes.
- This post describes a solution for blue/green Kubernetes cluster upgrades when Amazon EKS-A is deployed on vSphere using the bundled Flux controller.
- Windows Authentication on Amazon EKS Windows pods
- This post explains how to configure Windows containers to authenticate to Active Directory using gMSA.
- Karpenter for Kubernetes | Karpenter vs Cluster Autoscaler
- Karpenter consolidation in Kubernetes
- Enhancing Istio Ambient Mesh with eBPF
- Trying ValidationAdmissionPolicy / CEL Admission Policy from k8s 1.26
- Resize Your Pods In-Place with eBPF Triggers | KubeCon Detroit 2022
- KubeCon EU 2023 is now open for registration
- NVIDIA Kubernetes Device Plug-in Brings Temporal GPU Concurrency
- CNCF Tech for Self-Service Developer Platforms
- Tracing Kubernetes Adoption, From Inception to the Cloud
- Amazon EKS add-ons: Custom and advanced configuration with Terraform
- On Amazon EKS and SaaS
- hardeneks, a best practices analyzer for EKS
- eks-node-viewer, a tool for visualizing dynamic node usage within a cluster
- kubeshark, an API traffic viewer for Kubernetes
- kptop, a tool that displays Prometheus metrics from the command line
- KEDAlyze, visualizations for monitoring KEDA
- Launchpad by Jetstack, a command-line tool that lets you easily create applications on Kubernetes