KubeBlocks — Unified Database Operator for Kubernetes

Introduction

KubeBlocks provides a single operator that manages multiple database systems on Kubernetes. Instead of installing separate operators for MySQL, PostgreSQL, MongoDB, Redis, and messaging systems, KubeBlocks unifies provisioning, scaling, backup, and monitoring under one consistent API and CLI.

What KubeBlocks Does

• Provisions and manages MySQL, PostgreSQL, MongoDB, Redis, Kafka, and more via a single operator
• Automates backups, point-in-time recovery, and cross-region replication
• Handles vertical and horizontal scaling with automated failover
• Provides a unified CLI (kbcli) and CRD API for all supported database engines
• Supports add-ons to extend the list of managed database engines

Architecture Overview

KubeBlocks runs a set of controllers that watch Cluster CRDs. Each Cluster references a ClusterDefinition (the engine type) and a ClusterVersion (the specific version). The controllers create StatefulSets, Services, ConfigMaps, and backup jobs based on the definition. An add-on system lets the community contribute new engine definitions without modifying the core operator.

Self-Hosting & Configuration

• Install the operator via kbcli, Helm chart, or kubectl with the provided manifests
• Requires a Kubernetes 1.22+ cluster with a default StorageClass for persistent volumes
• Create clusters with kbcli cluster create specifying the engine, replicas, and resource limits
• Configure backups by creating a BackupPolicy resource pointing to an S3-compatible bucket
• Monitor databases via built-in Prometheus metrics exporters and optional Grafana dashboards

Key Features

• Single operator supporting 30+ database and middleware engines
• Declarative backup and restore with scheduled and on-demand options
• Automated vertical and horizontal scaling with connection draining
• Switchover and failover orchestration with configurable health probes
• Add-on architecture for contributing custom engine definitions

Comparison with Similar Tools

• CloudNativePG — PostgreSQL-only operator; KubeBlocks supports many engines
• Percona Operators — One operator per database; KubeBlocks unifies them
• Zalando Postgres Operator — PostgreSQL-specific; KubeBlocks covers MySQL, MongoDB, Redis, and more
• Sealos — Cloud platform that can use KubeBlocks for its database-as-a-service layer

FAQ

Q: Which databases does KubeBlocks support? A: MySQL, PostgreSQL, MongoDB, Redis, Kafka, Elasticsearch, RabbitMQ, and many more via add-ons.

Q: Does KubeBlocks replace my existing database operator? A: It can. KubeBlocks provides the same lifecycle management with a unified API across engines.

Q: Can I run KubeBlocks on managed Kubernetes services? A: Yes. KubeBlocks runs on EKS, GKE, AKS, and any conformant Kubernetes cluster.

Q: Is KubeBlocks free? A: Yes. KubeBlocks is open source under the GNU Affero General Public License v3.0.

Sources

• https://github.com/apecloud/kubeblocks
• https://kubeblocks.io/docs