EKS News 017

If you’re in Valencia this week for KubeCon + CloudNativeCon Europe 2022, please stop by the AWS booth and say hello. We love having conversations with y’all.

In this issue we’ll learn how to monitor EKS Anywhere using Amazon Managed Service for Prometheus and Amazon Managed Grafana, load testing with Locust, AWS Container Days, Kubernetes 1.24 significant changes, and more.

Monitor your Amazon Managed Service for Prometheus usage with Amazon CloudWatch usage metrics

Amazon Managed Service for Prometheus is generally available in the following regions: US East (Ohio), US East (N. Virginia), US West (Oregon), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), Europe (Stockholm), and Europe (London) Regions. To get started with usage metrics, check out our user guide on Amazon Managed Service for Prometheus’s CloudWatch usage metrics.

Amazon Managed Service for Prometheus is now available in Europe (London) Region

Amazon Managed Service for Prometheus is now generally available in the following regions: US East (Ohio), US East (N. Virginia), US West (Oregon), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), Europe (Stockholm), and Europe (London) Regions. To get started, check out the user guide, product page and pricing page for more information.

Monitoring Amazon EKS Anywhere using Amazon Managed Service for Prometheus and Amazon Managed Grafana

  • This blog provides a step-by-step guide on how to monitor your containerized workload running on Amazon EKS Anywhere by publishing metrics to Amazon Managed Service for Prometheus and using Amazon Managed Grafana to visualize
  • This is a complete walkthrough of the entire process; complete with YAML, commands, screen shots, etc.

Seamlessly migrate workloads from EKS self-managed node group to EKS-managed node groups

  • Briefly explains the benefits of MNGs: orchestrated upgrades, graceful node termination, and AMI management
  • Uses taints rather than cordon to prevent the scheduler from scheduling workloads on to nodes in the self managed node group
  • Involves increasing the number replicas for each deployment causing new pods to be scheduled onto nodes in the managed node group (use PDBs instead)
  • Once replicas are running on the MNG nodes, the self managed node group is deleted

Load testing your workload running on Amazon EKS with Locust

  • Locust (https://locust.io/) is an open-source load testing tool that comes with a real-time dashboard and programmable test scenarios
  • Walks through how to run Locust against workload running in an EKS cluster
  • Locust offers a nice simple web UI that measures number of users, response time, RPS, and other statistics

The Intel®3D Athlete Tracking (3DAT) scalable architecture deploys pose estimation models using Amazon Kinesis Data Streams and Amazon EKS

  • We explore the AWS services used to meet the 3DAT design requirements, including Amazon Kinesis Data Streams and Amazon Elastic Kubernetes Service (Amazon EKS), in order to scalably deploy the necessary pose estimation models for this software as a service (SaaS) application
  • A design choice was made to use Amazon EKS to host ML endpoints, giving the flexibility of running upstream Kubernetes with the option of clusters both fully managed in AWS via AWS Fargate, or on-premises hardware via Amazon EKS Anywhere
  • Each of the three models were packaged up into Docker containers—model files were stored in Amazon S3 and model images were stored in Amazon Elastic Container Registry (Amazon ECR)—and deployed as Cortex Realtime APIs

Join us for AWS Containers Days @ KubeCon + CloudNativeCon Europe 2022! In the days leading up to Kubecon + CloudNativeCon we’ll be spending the week talking about how you can deploy, manage, and scale containerized applications using Kubernetes on AWS. These sessions, led by our AWS Kubernetes team and guest speakers, will contain technical deep dives, product demos, and best practices.

We’re into day 4 already (time flies). If you want to catch up, here are the previous days videos on Twitch:

You aren’t Doing GitOps without Drift Detection

  • GitOps unifies ​​Kubernetes YAML files and application code in Git repositories to prevent configuration drift
  • ‘Configuration drift’ is a common term to describe this change that takes place in production environments (think of it like corrosion)
  • GitOps leverages an agent like Flux that is connected to all Git repositories in a system and owns the deployment of changes from these systems into production

Kubernetes 1.24: Stargazer

  • Dockershim Removed from kubelet
  • Beta APIs Off by Default
  • Signing Release Artifacts
  • Storage Capacity and Volume Expansion Are Generally Available
  • Storage Plugin Migration
  • CSI graduates to stable
  • and so much more!

sustainable-computing-io/kepler

  • Kepler (Kubernetes-based Efficient Power Level Exporter) uses eBPF to probe energy related system stats and exports as Prometheus metrics
  • I mentioned this project during our Container Day 3 session with Liz Rice and Duffie Cooley from Isovalent as a very unique way to use eBPF
  • This is a new open source project; it probably needs to be tested thoroughly before use in production
  • If you end up using the project, I’m very curious to hear how it’s going

Migrating from Cluster Autoscaler

  • This guide will show you how to switch from the Kubernetes Cluster Autoscaler to Karpenter for automatic node provisioning
  • All the information you need to go all-in on Karpenter
  • You literally end up removing the Cluster Autoscaler as part of the process

Get good at Git

  • No one really teaches us git; we have to learn the hard way
  • Could you be using Git more efficiently? The answer is, probably, a YES
  • In this blog, we will look at a bunch of different tips and tricks to make you more productive with Git