HANDS-ON OPTIMIZATION WORKSHOP

K8s Optimization Workshop — Why Your Nodes Are Half Empty (And How to Fix It)

Tuesday, May 19, 2026, 12:00pm ET

Your Kubernetes cluster is supposed to be an efficient scheduling machine. Instead, it's running half-empty nodes, over-provisioned workloads, and a cloud bill that makes finance uncomfortable. The culprit? Suboptimal binpacking.

Inefficient binpacking triggers a cascade of problems: nodes that appear busy but aren't, autoscalers firing at the wrong times, and a scheduler working with inaccurate inputs. The good news: most issues trace back to a handful of fixable misconfigurations.

In this hands-on workshop, Eli Birger and Anton Weiss walk you through the most common configuration mistakes behind poor node utilization — and show you how PerfectScale helps you identify and resolve them fast, with real examples you can follow along with.

What We'll Cover

Why your nodes aren't utilized efficiently and how inaccurate resource requests are usually the root cause
How misconfigured affinity rules, topology constraints, and priority classes create scheduling bottlenecks
Hands-on fixes that produce immediate improvements in node utilization
How PerfectScale surfaces binpacking inefficiencies and helps you resolve them before they compound

Who Should Attend

Platform and DevOps engineers who suspect their clusters are over-provisioned
SREs dealing with unpredictable autoscaling behavior
Anyone responsible for Kubernetes cost efficiency and cluster health

Eli Birger

Co-Founder and CTO of PerfectScale by DoiT

Eli is a passionate technologist with a background in Telecom (Comverse, Vonage), Cyber (Imperva, Cyren), and Storage (IBM). He has over six years of experience as a DevOps Manager, where he specialized in building large-scale SaaS systems based on Kubernetes. His current focus as PerfectScales CTO is on building solutions that solve the pains of operating highly-disturbed Kubernetes systems, keeping them running both efficiently and resiliently.