Karpenter: Can’t Live With It, Can’t Live Without It
- Monday June 22nd, 2026
- 4:00 PM CET / 10:00 AM EST
- Online
u003ch3u003eWebinar Detailsu003c/h3u003eu003culu003e
Karpenter changed the game for Kubernetes autoscaling. It’s fast, flexible, and the de facto choice for teams who outgrew Cluster Autoscaler. But once you run it in production at scale, the cracks show: nodes that won’t drain, consolidation that stalls, costs that creep back up no matter how many node pools you tune.
This isn’t a Karpenter teardown. Karpenter is doing exactly what it was designed to do — reactively optimize for the state of the cluster as it exists right now. The problem is that the state of the cluster is shaped by unevictable pods, scattered PDBs, and short-lived workloads that pin long-lived nodes. Reactive optimization can only do so much when the world it’s reacting to is fragmented.
Join Adi and Or for a candid technical session on where Karpenter hits a ceiling, and what to do about it.
Register Now
The Webinar Will Cover
How Karpenter Become The Standard
Why Karpenter wins where Cluster Autoscaler lost, and the design choices that come with that win
Operating Karpenter
What “properly tuned Karpenter” actually looks like, and where no amount of tuning will help. We will cover what good Karpenter operation looks like, PDBs, NodePools, and the discipline behind them
What’s Next?
Where this is heading: from reactive optimization to proactive placement. How to avoid meeting the issues before they hit hard.
