<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Kueue on vsz@ | Kubernetes, AI, tech stuff</title><link>https://victorszalvay.com/tags/kueue/</link><description>Recent content in Kueue on vsz@ | Kubernetes, AI, tech stuff</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>Victor Szalvay — All rights reserved</copyright><lastBuildDate>Sun, 16 Aug 2026 09:00:00 -0700</lastBuildDate><atom:link href="https://victorszalvay.com/tags/kueue/index.xml" rel="self" type="application/rss+xml"/><item><title>From Karpenter to GKE ComputeClasses, Part 3: accelerators, reservations, and the portability question</title><link>https://victorszalvay.com/karpenter-to-computeclasses-accelerators/</link><pubDate>Sun, 16 Aug 2026 09:00:00 -0700</pubDate><guid>https://victorszalvay.com/karpenter-to-computeclasses-accelerators/</guid><description>&lt;p&gt;&lt;em&gt;Part 3 of a 3-part series on porting EKS/Karpenter autoscaling patterns to GKE.
&lt;a href="https://victorszalvay.com/karpenter-to-computeclasses/" &gt;Part 1&lt;/a&gt; covered the mental model — a solver searching a
space versus an author ranking a list. &lt;a href="https://victorszalvay.com/karpenter-to-computeclasses-diversity/" &gt;Part 2&lt;/a&gt;
covered diversity, consolidation, and the where/when split between ComputeClass and
WorkloadClass. This one is the expensive end of the fleet, plus the promised answer on
portability.&lt;/em&gt;&lt;/p&gt;</description><media:content xmlns:media="http://search.yahoo.com/mrss/" url="https://victorszalvay.com/karpenter-to-computeclasses-accelerators/cover.png"/></item></channel></rss>