aws/karpenter-provider-aws released v1.15.0 #ec2 #karpenter #eks #kubernetes

Control Karpenter node termination with granular disruption budgets. Cron schedules + node limits for safe HA maintenance. Kubernetes 1.26+, EKS, Karpenter https://www.valtersit.com/vault/setting-granular-disruption-budgets-for-highavailability-mai-c8e9b8/

#karpenter #high-availability #ValtersIT

https://lckhd.eu/Dtyw3w

#Karpenter #EKS #SpotInstances #Kubernetes #FinOps

There are many ways to optimize EC2 and EKS spend on AWS. One approach is using Spot instances, which AWS can reclaim on short notice.

https://lckhd.eu/c22cFY

#Karpenter #EKS #Graviton #Kubernetes #FinOps

One way to save costs on AWS is to run on Graviton (ARM) compute. You will usually save money compared to x86 and in many cases get better performance. On EKS you can run ARM and x86 worker nodes side by side, but your build

Stop hardcoding subnet and SG IDs in EC2NodeClass. Use AWS tags instead: Karpenter queries the EC2 API at runtime, picks subnets by free IP space, and matches security https://www.valtersit.com/vault/configuring-ec2nodeclass-for-dynamic-subnet-and-security-gro-f207e0/ #karpenter #aws #ValtersIT

Karpenter multi-AZ Spot pools with on-demand fallback: diverse instance families, lowest-price selection, minimal disruption. Full config:
https://www.valtersit.com/vault/provisioning-flexible-multiaz-spot-instance-pools-with-ondem-4897d8/
#karpenter #spot-instances #aws

ReadyOn’s Four Walls of tenant isolation on Amazon EKS #karpenter #eks #kubernetes

✍️ New blog post by Frank Osasere Idugboe

Kubernetes at 3% CPU but Pods Still Pending: Debugging Requests, Karpenter, and EBS AZ Affinity

#aws #kubernetes #karpenter #devops

https://lckhd.eu/bFdnju

#Kubernetes #Minikube #EKS #Karpenter #ArgoCD

Many Kubernetes tutorials stay at the high level, which is understandable given how much ground there is to cover. The article below goes into more detail, taking you from a fresh Minikube install through a working GitOps loop

https://lckhd.eu/XD1fLf

#Karpenter #EKS #Kubernetes #Terraform #AWS

The Cluster Autoscaler (CAS) scales EKS worker nodes by adding instances to Auto Scaling Groups (ASGs), and each group knows one instance type. In practice that means a node group per instance type per AZ, and pods sitting