<?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>Nftables on KubernetesLab</title><link>https://kuberneteslab.dev/en/tags/nftables/</link><description>Recent content in Nftables on KubernetesLab</description><image><title>KubernetesLab</title><url>https://kuberneteslab.dev/images/og-brand.png</url><link>https://kuberneteslab.dev/images/og-brand.png</link></image><generator>Hugo</generator><language>en</language><lastBuildDate>Tue, 04 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://kuberneteslab.dev/en/tags/nftables/index.xml" rel="self" type="application/rss+xml"/><item><title>How much CPU and memory does a CNI actually use?</title><link>https://kuberneteslab.dev/en/blog/cni-standing-cost/</link><pubDate>Tue, 04 Aug 2026 00:00:00 +0000</pubDate><guid>https://kuberneteslab.dev/en/blog/cni-standing-cost/</guid><description>We measured the standing CPU and memory usage of Calico, Cilium, Flannel, Antrea, and kube-router across 14 configurations in a 9-day unattended campaign. Memory, not CPU, is what separates them, and switching kube-proxy to nftables mode alone cut its memory usage by 70%.</description></item></channel></rss>