<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Cloud-Native on Taubyte Blog</title><link>/blog/tags/cloud-native/</link><description>Recent content in Cloud-Native on Taubyte Blog</description><image><title>Taubyte Blog</title><url>/blog/opengraph.jpg</url><link>/blog/opengraph.jpg</link></image><generator>Hugo -- 0.146.0</generator><language>en-us</language><lastBuildDate>Sun, 15 Feb 2026 16:00:00 +0000</lastBuildDate><atom:link href="/blog/tags/cloud-native/index.xml" rel="self" type="application/rss+xml"/><item><title>Kubernetes: The Orchestration Tax Most Teams Don't Need</title><link>/blog/posts/kubernetes-performance-fresser/</link><pubDate>Sun, 15 Feb 2026 16:00:00 +0000</pubDate><guid>/blog/posts/kubernetes-performance-fresser/</guid><description>&lt;p>Kubernetes was built to orchestrate Google&amp;rsquo;s global infrastructure. You are not Google. Terribly sorry.&lt;/p>
&lt;p>82% of container users run Kubernetes in production. Most of them shouldn&amp;rsquo;t.&lt;/p>
&lt;h2 id="the-control-plane-tax">The Control Plane Tax&lt;/h2>
&lt;p>Before your application serves a single request, Kubernetes needs etcd chewing through 2-8 GB RAM per node. Then kube-apiserver, kube-scheduler, kube-controller-manager, kubelet (reserving 25% of node memory by default), CoreDNS, kube-proxy, and a CNI plugin. All of this before your code runs.&lt;/p></description></item></channel></rss>