Book Image

Kubernetes in Production Best Practices

By : Aly Saleh, Murat Karslioglu
Book Image

Kubernetes in Production Best Practices

By: Aly Saleh, Murat Karslioglu

Overview of this book

Although out-of-the-box solutions can help you to get a cluster up and running quickly, running a Kubernetes cluster that is optimized for production workloads is a challenge, especially for users with basic or intermediate knowledge. With detailed coverage of cloud industry standards and best practices for achieving scalability, availability, operational excellence, and cost optimization, this Kubernetes book is a blueprint for managing applications and services in production. You'll discover the most common way to deploy and operate Kubernetes clusters, which is to use a public cloud-managed service from AWS, Azure, or Google Cloud Platform (GCP). This book explores Amazon Elastic Kubernetes Service (Amazon EKS), the AWS-managed version of Kubernetes, for working through practical exercises. As you get to grips with implementation details specific to AWS and EKS, you'll understand the design concepts, implementation best practices, and configuration applicable to other cloud-managed services. Throughout the book, you’ll also discover standard and cloud-agnostic tools, such as Terraform and Ansible, for provisioning and configuring infrastructure. By the end of this book, you’ll be able to leverage Kubernetes to operate and manage your production environments confidently.
Table of Contents (12 chapters)

Configuring Kube Proxy

kube-proxy is an agent service that runs on each node in the cluster to create, update, and delete network rules on the nodes, usually through the use of Linux iptables. These network rules allow inter-pod and intra-pod communication inside and outside the Kubernetes cluster.

Irrespective of whether you use a self-managed Kubernetes cluster or a hosted one, you need to control the configuration options that you pass to kube-proxy. As we are using EKS, kube-proxy comes pre-deployed with the cluster, which leaves us without a full control over its configuration, and we need to change this.

During the cluster's lifetime, you need to control the periodic updates of kube-proxy and include them within the cluster's updates' pipeline. Also, you need to optimize its performance by controlling the runtime parameters, including --iptables-sync-period, --iptables-min-sync-period, and --proxy-mode.

To learn about the remainder of the configuration...