DevOps Daily with Fexingo · 2026-08-20 · 8 min
In this episode of DevOps Daily, Lucas and Luna dig into a deceptively simple failure mode: the Kubernetes Metrics Server takes a few extra seconds to return pod metrics, and the Horizontal Pod Autoscaler overreacts. They walk through a real incident where a 45-second metrics delay caused wild replica swings, wasted cloud spend, and a brief outage. Lucas explains how the HPA's default tolerance window interacts with the Metrics Server's 15-second scrape interval and why a simple misconfiguration can turn a steady workload into a yo-yo. They discuss how to monitor metrics-server latency, the role of the - horizontal-pod-autoscaler-sync-period flag, and why tuning your autoscaler's behavior to actual metrics freshness is more important than chasing the latest feature. Along the way, they tie the lesson to the broader principle: in Kubernetes, the control plane's speed dictates your application's stability. A short, honest word about listener support keeps the show ad-free, then it's back to the metrics trench.
How Kubernetes Namespace Quotas Trigger Unexpected Failures58 / 100Why Your Kubernetes Cost Reports Are Lying To YouHow Kubernetes Node Pools Can Cost You More Than You ThinkHow Kubernetes Pod Overhead Changes Node Capacity CalculationsHow Kubernetes Pod Security Admission Blocks Risky Workloads