Radosław from aleno migrated from ECS to Kubernetes — spot instances broke at 20+ containers, firewalls silently dropped traffic, and the wrong memory metric caused OOM kills.

You will learn:

- Why ECS spot instances have no built-in fallback mechanism — and how Karpenter's flexible instance selection solves it
- How running Flux and Argo CD together gives infra teams git-push workflows while developers get a UI
- Why the default Kubernetes memory metric includes evictable caches — and switching to working set fixed OOM errors
- How jemalloc cut memory usage by 20% and fixed HPA autoscaling for WebSocket containers

Watch (or listen to) it here: https://ku.bz/x6wFMhVsx

🌟 This episode is sponsored by LearnKube — get started on your Kubernetes journey through comprehensive online, in-person, or remote training https://learnkube.com/training

With @Birthmarkb