InITNEXTbySebastian Daberdaku·19h agoWhat ten months of deleting idle nodes taught us about KarpenterWe wrote a small tool to remove Kubernetes nodes our autoscaler left behind. Its logs showed that most of its work belonged in Karpenter’s…
Chinmay Tonape·5d agoScaling an EKS Cluster with KarpenterWhen a Kubernetes cluster runs out of room, something has to add nodes. For years that job belonged to the Cluster Autoscaler, which scales…A response icon1A response icon1
Naveena Ravi·Oct 27Implementation of AWS EKS for AI applicationsAWS EKS Implementation with Karpenter, Amazon EKS managed node group, Amazon managed Prometheus (AMP), Amazon managed Grafana and AWS…
Anish Bhat K·Oct 26Karpenter Explained: Why Not Just Use Cluster Autoscaler?When working with Kubernetes, one thing we need to get right is how the cluster scales when more workloads arrive.
InITNEXTbyAhmad Babaei·Oct 23Karpenter’s Chicken-and-Egg Problem on EKSFlux installed Karpenter, and Flux needed a node from Karpenter. Moving the bootstrap into Terraform broke the loop, and CoreDNS hid a…A response icon1A response icon1
InITNEXTbySebastian Daberdaku·Oct 17Running Trino workers on spot instancesGraceful shutdown, Karpenter node pools on local NVMe, spill to disk, and fault-tolerant execution spooling to S3 Express One Zone.
Arantika Bagewadi·Oct 13Karpenter v1 Is a Two-Hop Upgrade, and the Second Hop Is the Dangerous OnePart 1 of 4 on migrating Karpenter from v1beta1 to v1 without breaking prod. This series is a reference walkthrough built from Karpenter’s…
Kirubaashini Selvakumar·Oct 5Same Image, Cheaper Nodes: Multi Platform Docker Builds for Karpenter on GravitonBuild once for amd64 and arm64 with docker buildx, then let Karpenter pick cheaper Graviton nodes without exec format errors
Mridultiwari·Oct 4When the cluster looked idle but nothing would scheduleThe first thing I checked was the obvious one: CPU. Our non-prod EKS workers are Karpenter-managed, and application pods had been sitting…
Ayoub Reda·Oct 3Running ARM and x86 Side by Side on EKS with KarpenterHow we built every image twice, moved the cluster onto Graviton and Bottlerocket, and let Karpenter pick the cheapest CPU for each node
