Deploy production-grade RabbitMQ clusters on Kubernetes in minutes. Automate HA with quorum queues (Raft replication), and zero-downtime upgrades.
Leader Re-election
Open Source
Deploy RabbitMQ 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 RabbitMQ Addon
helm upgrade -i kb-addon-rabbitmq kubeblocks/rabbitmq \ --namespace kb-system
Create a RabbitMQ Cluster
apiVersion: apps.kubeblocks.io/v1
kind: Cluster
metadata:
name: rabbitmq-cluster
namespace: demo
spec:
terminationPolicy: Delete
componentSpecs:
- name: rabbitmq
componentDef: rabbitmq
serviceVersion: "3.13.7"
replicas: 3Cluster is Ready
$ kubectl get cluster rabbitmq-cluster -n demo NAME CLUSTER-DEFINITION TERMINATION-POLICY STATUS AGE rabbitmq-cluster Delete 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 GridBONC CloudChina Mobile CloudChina Telecom CloudTencentXiaomiPing AnVIP.comKwaiTiger BrokersCITIC SecuritiesSealOSFastGPTKubeSphereMomenta360TilaaOlaresChangan AutomobileState GridKubeBlocks deploys RabbitMQ as a 3- or 5-node HA cluster. Quorum queues use Raft replication for strong consistency; the cluster itself uses Erlang distributed clustering with Kubernetes peer discovery.
KubeBlocks deploys RabbitMQ as a 3-node (or 5-node) HA cluster. Quorum queues replicate every message to a majority of nodes via Raft before acknowledging the producer. If a queue leader fails, the remaining members elect a new one automatically — no manual intervention required.
Raft consensus — every quorum queue write is replicated to a majority before acknowledged
Automatic leader re-election in under 30 seconds on node failure
All nodes accept AMQP connections — clients reconnect to any surviving node
Peer discovery via Kubernetes Endpoints API (no external etcd or ZooKeeper)
Classic queues and quorum queues supported side-by-side
Management UI on :15672, Prometheus metrics on :15692 per node
No SSH into pods, no shell scripts. Submit an OpsRequest and KubeBlocks handles the rest.
Availability & Scaling
Horizontal Scaling
Add or remove nodes (3→5 or 5→3) online. KubeBlocks joins new pods to the RabbitMQ cluster and rebalances queue leaders.
Vertical Scaling
Resize CPU and memory on running nodes with a rolling strategy and no message loss.
Volume Expansion
Expand PVC storage for message data without pod restarts on supported storage classes.
Rolling Restart
Controlled pod restarts that maintain quorum throughout — the cluster stays available.
Stop / Start
Suspend the cluster to eliminate compute cost; resume with full state and quorum intact.
Configuration, Security & Observability
Dynamic Configuration
Apply RabbitMQ parameters via OpsRequest. KubeBlocks applies changes with a rolling pod restart, keeping quorum available throughout.
Credential Management
Admin credentials are auto-generated and stored in Kubernetes Secrets at cluster creation. Update via Secret rotation.
Version Upgrade
Rolling upgrades across supported versions (e.g. 3.13.7 → 4.0.9) with health checks between each pod.
Prometheus Metrics
Per-node metrics on :15692 via the built-in Prometheus plugin. Grafana dashboards available.
Expose via LoadBalancer
Expose AMQP or Management UI externally via a LoadBalancer or NodePort service.
KubeBlocks automates the hardest parts of running RabbitMQ on Kubernetes — so your team doesn't have to.
Deploy a production-grade RabbitMQ cluster in minutes with HA, quorum queues (Raft replication), and full Day-2 operations — all open source.
