How do I deploy etcd on Kubernetes using KubeBlocks?

Install KubeBlocks via Helm, then install the etcd addon: "helm upgrade -i kb-addon-etcd kubeblocks/etcd -n kb-system". Create an etcd cluster by applying a Cluster resource with clusterDef: etcd, your desired replica count (3 or 5), and serviceVersion. KubeBlocks provisions pods, services, and PVCs automatically.

What etcd versions does KubeBlocks support?

KubeBlocks supports etcd 3.5.6, 3.5.15, and 3.6.1. Rolling upgrades between versions are supported via OpsRequest.

How does KubeBlocks handle etcd high availability?

KubeBlocks deploys etcd in a Raft cluster with 3 or 5 nodes. A 3-node cluster tolerates 1 failure; a 5-node cluster tolerates 2. When the leader fails, Raft elects a new leader among the remaining members within seconds — no manual intervention needed.

Does KubeBlocks etcd Operator support backup and restore?

Yes. KubeBlocks supports snapshot-based backup for etcd using etcdctl to create consistent DB snapshots, which are then uploaded to S3-compatible object storage. Scheduled backups are supported via the BackupSchedule CRD. Restore creates a new cluster from any stored snapshot.

Can I scale my etcd cluster horizontally with KubeBlocks?

Yes. KubeBlocks supports horizontal scaling of etcd clusters via OpsRequest. You can safely grow from 3 to 5 nodes — KubeBlocks joins new pods to the Raft cluster, ensuring quorum is maintained at every step. Scaling from 5 to 3 is also supported.

Is the KubeBlocks etcd Operator open source?

Yes. KubeBlocks is open source under the AGPL-3.0 license at https://github.com/apecloud/kubeblocks. The etcd addon is available at https://github.com/apecloud/kubeblocks-addons. An Enterprise edition adds a web management UI and dedicated support.

The Production-Grade etcd Operator for Kubernetes | KubeBlocks

Open Source · CNCF Graduated · Powers Kubernetes

KubeBlocks etcd Operator for Kubernetes

Deploy production-grade etcd clusters in minutes. Raft HA, snapshot backup, horizontal scaling, and full Day-2 operations via a single operator.

Try Playground Free →Read the Docs

< 5s

Leader Election

99.99%

HA Availability

100%

Open Source

Deploy etcd in 4 steps

Install KubeBlocks

# Add Helm repo
helm repo add kubeblocks https://apecloud.github.io/helm-charts
helm repo update

# Install KubeBlocks
helm install kubeblocks kubeblocks/kubeblocks \
  --namespace kb-system --create-namespace

Install etcd Addon

helm upgrade -i kb-addon-etcd kubeblocks/etcd \
  -n kb-system

Create an etcd Cluster

apiVersion: apps.kubeblocks.io/v1
kind: Cluster
metadata:
  name: etcd-cluster
  namespace: demo
spec:
  terminationPolicy: Delete
  clusterDef: etcd
  componentSpecs:
    - name: etcd
      serviceVersion: "3.5.15"
      replicas: 3

Cluster is Ready

$ kubectl get cluster etcd-cluster -n demo
NAME           CLUSTER-DEF   STATUS    AGE
etcd-cluster   etcd          Running   2m

Trusted by Engineering Teams at Scale

BONC Cloud

China Mobile Cloud

China Telecom Cloud

Tencent

Xiaomi

Ping An

VIP.com

Kwai

Tiger Brokers

CITIC Securities

SealOS

FastGPT

KubeSphere

Momenta

360

Tilaa

Olares

Changan Automobile

State Grid

BONC Cloud

China Mobile Cloud

China Telecom Cloud

Tencent

Xiaomi

Ping An

VIP.com

Kwai

Tiger Brokers

CITIC Securities

SealOS

FastGPT

KubeSphere

Momenta

360

Tilaa

Olares

Changan Automobile

State Grid

Architecture

Raft Consensus. One Operator.

One etcd Raft HA topology — leader, followers, and quorum — operated consistently on Kubernetes.

KubeBlocks runs etcd as one Raft HA architecture on Kubernetes: one leader handles writes, followers replicate the WAL, and a majority of members must acknowledge before a write is committed. Cluster size is configurable for your SLOs; the same operator workflow covers backup, restore, switchover, TLS, and Day-2 ops.

✓

Majority quorum — committed writes are replicated to most members before acknowledgment

✓

Automatic leader election when the current leader is unavailable

✓

Headless service for stable per-pod DNS; optional ClusterIP client service

✓

Consistent snapshot backup to object storage (full snapshot, not PITR)

Raft consensus

Majority quorum

Automatic failover

Snapshot backup

Application / Client

Client (if enabled) {cluster}-etcd-client:2379
Default {cluster}-etcd-headless (pod DNS)

client API → all pods (etcd routes internally)

Kubernetes Services

{cluster}-etcd-client

ClusterIP · :2379 client
all pods (no roleSelector)
disableAutoProvision: true — not created by default

Optional

→ any pod (etcd forwards to leader transparently)

Pods · Worker Nodes

etcd-0LEADER

🔑

etcd

:2379 client + /metrics · :2380 peer

leader

💾 PVC data-0 · 10Gi

etcd-1FOLLOWER

🔑

etcd

:2379 client + /metrics · :2380 peer

💾 PVC data-1 · 10Gi

etcd-2FOLLOWER

🔑

etcd

:2379 client + /metrics · :2380 peer

💾 PVC data-2 · 10Gi

↔Raft ConsensusWAL replicated to followers · quorum acknowledgment required

Leader Node

Follower Node

Persistent Storage

Comparison

How KubeBlocks Compares to Other etcd Options

Feature	KubeBlocks	Bitnami Helm Chart	etcd-operator
Kubernetes-native CRD API	✓	✗	✓
Raft HA (3 / 5-node)	✓	✓	✓
TLS encryption (peer & client)	✓	✓	~
Leader switchover (Switchover)	✓	✗	~
Horizontal scaling	✓	✓	✓
Vertical scaling (CPU/memory)	✓	✓	✗
PVC volume expansion	✓	✗	✗
Parameter reconfiguration	✓	✓	✗
Rolling version upgrade	✓	✓	✗
Snapshot backup & restore	✓	~	~
Prometheus metrics	✓	✓	✓
Stop / start cluster	✓	✗	✗
Open Source	✓	✓	✓
Cluster management web UI	Enterprise	✗	✗

✓ = Supported · ~ = Partial / Limited · ✗ = Not supported

Enterpriseindicates a capability available in KubeBlocks Enterprise, not the open-source distribution. Contact us for licensing →

etcd-operator column based on hands-on testing of etcd-io/etcd-operator v0.2.0 (Mar 2026) on EKS. Other columns based on official documentation and GitHub issues. Features may vary by version.

Capabilities

Built for Production etcd

KubeBlocks automates the hardest parts of running etcd on Kubernetes — so your team doesn't have to.

📸

Snapshot Backup & Restore

Scheduled snapshot backups via etcdctl to S3-compatible object storage. Restore a full cluster from any stored snapshot.

S3 Storage

Backup Timeline

Snapshot

00:00

Snapshot

+12h

Snapshot

+24h

Restore

target

📸

etcdctl snapshotConsistent DB snapshot

→

☁️

Upload to S3Durable object storage

→

🚀

Restore ClusterNew cluster from snapshot

✓ Stored Safely

Day-2 Operations

Every Operation Declared as a Kubernetes Resource

No SSH into pods, no shell scripts. Submit an OpsRequest and KubeBlocks handles the rest.

Scaling & Availability

✓

Horizontal Scaling

Add or remove etcd members (3 → 5 or 5 → 3) online. Raft safely admits new members and redistributes the quorum.

✓

Vertical Scaling

Resize CPU and memory on etcd pods with a rolling strategy that keeps quorum available throughout.

✓

Volume Expansion

Expand etcd PVCs to increase data directory capacity without pod restarts.

✓

Rolling Restart

Controlled pod restarts one at a time — quorum is never lost during the operation.

✓

Stop / Start

Suspend clusters to eliminate compute cost during off-hours; resume with full cluster state.

Configuration, Security & Observability

✓

Dynamic Configuration

Tune etcd parameters such as heartbeat interval and election timeout via OpsRequest.

✓

Version Upgrade

Rolling upgrades from 3.5.x to 3.6.x with health checks. Quorum is maintained at each step.

✓

Prometheus Metrics

Built-in Prometheus metrics endpoint at :2379/metrics — plug into Grafana dashboards.

✓

Snapshot Backup & Restore

Consistent snapshots via etcdctl to S3-compatible object storage; restore from any stored snapshot.

✓

TLS Encryption

Enable mTLS for both client-server (:2379) and peer (:2380) communication with auto-managed certificates via cert-manager.

✓

Expose via LoadBalancer

Enable external client access for applications running outside the Kubernetes cluster.

From the Blog

Go Deeper on etcd on Kubernetes

We Let an AI Agent Manage Our Databases. Here's Why Most Operators Failed It.

We tested AI agents against traditional Kubernetes database operators — and why a unified API like KubeBlocks changes everything.

How to Manage Database Clusters Without a Dedicated Operator?

A unified approach to managing diverse databases on Kubernetes — including infrastructure services like etcd — without the overhead of dedicated operators.

Running Databases on Kubernetes — Insights from Leading Chinese Internet Companies

Why leading internet companies are increasingly adopting the practice of running stateful databases on Kubernetes.

Get Started

Get Started with KubeBlocks etcd Operator, Risk-Free.

Open source and production-ready. Enterprise customers get dedicated onboarding, migration support, and SLA guarantees.

Try Playground Free →Talk to the Team

✓Open Source

✓etcd & 35+ other engines

✓Production-grade HA

✓No vendor lock-in

# Add Helm repo helm repo add kubeblocks https://apecloud.github.io/helm-charts helm repo update # Install KubeBlocks helm install kubeblocks kubeblocks/kubeblocks \ --namespace kb-system --create-namespace

apiVersion: apps.kubeblocks.io/v1 kind: Cluster metadata: name: etcd-cluster namespace: demo spec: terminationPolicy: Delete clusterDef: etcd componentSpecs: - name: etcd serviceVersion: "3.5.15" replicas: 3

Feature

KubeBlocks

Bitnami Helm Chart

etcd-operator

Kubernetes-native CRD API

✓

✗

✓

Raft HA (3 / 5-node)

✓

TLS encryption (peer & client)

✓

Leader switchover (Switchover)

✓

✗

Horizontal scaling

✓

Vertical scaling (CPU/memory)

✓

✗

PVC volume expansion

✓

✗

Parameter reconfiguration

✓

✗

Rolling version upgrade

✓

✗

Snapshot backup & restore

✓

Prometheus metrics

✓

Stop / start cluster

✓

✗

Open Source

✓

Cluster management web UI

Enterprise

✗