# Why etcd breaks at scale in Kubernetes

DevFeed: [Why etcd breaks at scale in Kubernetes](<https://devfeed.tech/articles/why-etcd-breaks-at-scale-in-kubernetes-19130.md>)

Original publisher: [Read original article](<https://learnkube.com/etcd-breaks-at-scale>)

Author: Daniele Polencic

Published: 2026-02-24T00:00:00Z

Content type: tutorial

Language: en

Sources: [Learnk8s](<https://devfeed.tech/sources/learnk8s.md>)

Topics: [Kubernetes](<https://devfeed.tech/topics/kubernetes.md>), [Database](<https://devfeed.tech/topics/database.md>), [Raft](<https://devfeed.tech/topics/raft.md>), [Availability](<https://devfeed.tech/topics/availability.md>)

Tags: [api-server](<https://devfeed.tech/tags/api-server.md>), [consistency](<https://devfeed.tech/tags/consistency.md>), [database](<https://devfeed.tech/tags/database.md>), [high-availability](<https://devfeed.tech/tags/high-availability.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [raft](<https://devfeed.tech/tags/raft.md>), [scale](<https://devfeed.tech/tags/scale.md>)

## AI overview

This article explains why etcd becomes a bottleneck as Kubernetes clusters grow. It describes etcd as the API server's private, consistent backend and explains how Raft supports shared state across multiple API servers for high availability.

## Source excerpt

etcd is the database behind every Kubernetes cluster. This article explains etcd's design limits, how the API server amplifies them, and what you can do about it -- from sharding to what GKE, EKS, and k3s built.