# Sharding etcd on hosted control planes and what the numbers actually say

DevFeed: [Sharding etcd on hosted control planes and what the numbers actually say](<https://devfeed.tech/articles/sharding-etcd-on-hosted-control-planes-and-what-the-numbers-actually-say-85852.md>)

Original publisher: [Read original article](<https://developers.redhat.com/articles/2026/10/13/sharding-etcd-on-hosted-control-planes-and-what-the-numbers-actually-say>)

Author: Vishnu Challa

Published: 2026-10-13T13:17:03Z

Content type: article

Language: en

Sources: [Red Hat](<https://devfeed.tech/sources/red-hat.md>), [Red Hat Developer](<https://devfeed.tech/sources/red-hat-developer.md>)

Topics: [etcd](<https://devfeed.tech/topics/etcd.md>), [Distributed Systems & Parallel Computing](<https://devfeed.tech/topics/distributed-systems-parallel-computing.md>), [sharding](<https://devfeed.tech/topics/sharding.md>)

Tags: [capacity](<https://devfeed.tech/tags/capacity.md>), [capacity-planning](<https://devfeed.tech/tags/capacity-planning.md>), [clusters](<https://devfeed.tech/tags/clusters.md>), [containers](<https://devfeed.tech/tags/containers.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [pods](<https://devfeed.tech/tags/pods.md>)

## AI overview

The article reports benchmark results for etcd sharding in hosted Kubernetes control planes. In tests using two shards, separating event objects improved write throughput, compaction, and slow-apply results under event pressure, with comparable pod latency and workload completion times. Sharding used roughly 1.8 to 2.3 times the etcd container memory, and compaction was slower in a large custom-resource quota test. The authors caution that the feature is in tech preview and the results are not capacity-planning guidance.

## Source excerpt

Every write any workload makes to a Kubernetes cluster ends up in the same place: etcd's write-ahead log, and then its backend, and then eventually its compaction cycle. On a hosted control plane, that pressure is concentrated, and the control plane is a set of pods on a management cluster, while etcd is the component that decides how well it holds up under load. Much of that pressure comes from objects nothing reads. The post Sharding etcd on hosted control planes and what the numbers actually say appeared first on Red Hat Developer.