Valkey is an open source (BSD) high-performance key/value datastore that supports a variety workloads such as caching, message queues, and can act as a primary database.

Overview of Valkey Cluster

Trademarks: This software listing is packaged by Bitnami. The respective trademarks mentioned in the offering are owned by the respective companies, and use of them does not imply any affiliation or endorsement.

TL;DR

helm install my-release oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository.

Introduction

This chart bootstraps a Valkey deployment on a Kubernetes cluster using the Helm package manager.

Bitnami charts can be used with Kubeapps for deployment and management of Helm Charts in clusters.

Choose between Valkey Helm Chart and Valkey Cluster Helm Chart

You can choose any of the two Valkey Helm charts for deploying a Valkey cluster. While Valkey Helm Chart will deploy a primary-replica cluster using Valkey Sentinel, the Valkey Cluster Helm Chart will deploy a Valkey Cluster with sharding. The main features of each chart are the following:

Valkey	Valkey Cluster
Supports multiple databases	Supports only one database. Better if you have a big dataset
Single write point (single primary)	Multiple write points (multiple primary nodes)

Prerequisites

• Kubernetes 1.23+
• Helm 3.8.0+
• PV provisioner support in the underlying infrastructure

Installing the Chart

To install the chart with the release name my-release:

helm install my-release oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts.

The command deploys Valkey on the Kubernetes cluster in the default configuration. The Parameters section lists the parameters that can be configured during installation.

NOTE: if you get a timeout error waiting for the hook to complete increase the default timeout (300s) to a higher one, for example:

helm install --timeout 600s myrelease oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts. Tip: List all releases using helm list

Configuration and installation details

Resource requests and limits

Bitnami charts allow setting resource requests and limits for all containers inside the chart deployment. These are inside the resources value (check parameter table). Setting requests is essential for production workloads and these should be adapted to your specific use case.

To make this process easier, the chart contains the resourcesPreset values, which automatically sets the resources section according to different presets. Check these presets in the bitnami/common chart. However, in production workloads using resourcesPreset is discouraged as it may not fully adapt to your specific needs. Find more information on container resource management in the official Kubernetes documentation.

Update credentials

The Bitnami Valkey Cluster chart, when upgrading, reuses the secret previously rendered by the chart or the one specified in existingSecret. To update credentials, use one of the following:

• Run helm upgrade specifying a new password in password
• Run helm upgrade specifying a new secret in existingSecret

Prometheus metrics

This chart can be integrated with Prometheus by setting metrics.enabled to true. This will deploy a sidecar container with redis_exporter in all pods and a metrics service, which can be configured under the metrics.service section. This metrics service will have the necessary annotations to be automatically scraped by Prometheus.

Prometheus requirements

It is necessary to have a working installation of Prometheus or Prometheus Operator for the integration to work. Install the Bitnami Prometheus helm chart or the Bitnami Kube Prometheus helm chart to easily have a working Prometheus in your cluster.

Integration with Prometheus Operator

The chart can deploy ServiceMonitor objects for integration with Prometheus Operator installations. To do so, set the value metrics.serviceMonitor.enabled=true. Ensure that the Prometheus Operator CustomResourceDefinitions are installed in the cluster or it will fail with the following error:

no matches for kind "ServiceMonitor" in version "monitoring.coreos.com/v1"

Install the Bitnami Kube Prometheus helm chart for having the necessary CRDs and the Prometheus Operator.

Rolling VS Immutable tags

It is strongly recommended to use immutable tags in a production environment. This ensures your deployment does not change automatically if the same tag is updated with a different image.

Bitnami will release a new chart updating its containers if a new version of the main container, significant changes, or critical vulnerabilities exist.

Use a different Valkey version

To modify the application version used in this chart, specify a different version of the image using the image.tag parameter and/or a different repository using the image.repository parameter.

Cluster topology

To successfully set the cluster up, it will need to have at least 3 primary nodes. The total number of nodes is calculated like- nodes = numOfPrimaryNodes + numOfPrimaryNodes * replicas. Hence, the defaults cluster.nodes = 6 and cluster.replicas = 1 means, 3 primary and 3 replica nodes will be deployed by the chart.

By default the Valkey Cluster is not accessible from outside the Kubernetes cluster, to access the Valkey Cluster from outside you have to set cluster.externalAccess.enabled=true at deployment time. It will create in the first installation only 6 LoadBalancer services, one for each Valkey node, once you have the external IPs of each service you will need to perform an upgrade passing those IPs to the cluster.externalAccess.service.loadbalancerIP array.

The replicas will be read-only replicas of the primary nodes. By default only one service is exposed (when not using the external access mode). You will connect your client to the exposed service, regardless you need to read or write. When a write operation arrives to a replica it will redirect the client to the proper primary node. For example, using valkey-cli you will need to provide the -c flag for valkey-cli to follow the redirection automatically.

Using the external access mode, you can connect to any of the pods and the replicas will redirect the client in the same way as explained before, but the all the IPs will be public.

In case the primary crashes, one of his replicas will be promoted to primary. The slots stored by the crashed primary will be unavailable until the replica finish the promotion. If a primary and all his replicas crash, the cluster will be down until one of them is up again. To avoid downtime, it is possible to configure the number of Valkey nodes with cluster.nodes and the number of replicas that will be assigned to each primary with cluster.replicas. For example:

• cluster.nodes=9 ( 3 primary plus 2 replicas for each primary)
• cluster.replicas=2

Providing the values above, the cluster will have 3 primarys and, each primary, will have 2 replicas.

NOTE: By default cluster.init will be set to true in order to initialize the Valkey Cluster in the first installation. If for testing purposes you only want to deploy or upgrade the nodes but avoiding the creation of the cluster you can set cluster.init to false.

Adding a new node to the cluster

There is a job that will be executed using a post-upgrade hook that will allow you to add a new node. To use it, you should provide some parameters to the upgrade:

• Pass as password the password used in the installation time. If you did not provide a password follow the instructions from the NOTES.txt to get the generated password.
• Set the desired number of nodes at cluster.nodes.
• Set the number of current nodes at cluster.update.currentNumberOfNodes.
• Set to true cluster.update.addNodes.

The following will be an example to add one more node:

helm upgrade --timeout 600s <release> --set "password=${VALKEY_PASSWORD},cluster.nodes=7,cluster.update.addNodes=true,cluster.update.currentNumberOfNodes=6" oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts.

Where VALKEY_PASSWORD is the password obtained with the command that appears after the first installation of the Helm Chart. The cluster will continue up while restarting pods one by one as the quorum is not lost.

External Access

If you are using external access, to add a new node you will need to perform two upgrades. First upgrade the release to add a new Valkey node and to get a LoadBalancerIP service. For example:

helm upgrade <release> --set "password=${VALKEY_PASSWORD},cluster.externalAccess.enabled=true,cluster.externalAccess.service.type=LoadBalancer,cluster.externalAccess.service.loadBalancerIP[0]=<loadBalancerip-0>,cluster.externalAccess.service.loadBalancerIP[1]=<loadbalanacerip-1>,cluster.externalAccess.service.loadBalancerIP[2]=<loadbalancerip-2>,cluster.externalAccess.service.loadBalancerIP[3]=<loadbalancerip-3>,cluster.externalAccess.service.loadBalancerIP[4]=<loadbalancerip-4>,cluster.externalAccess.service.loadBalancerIP[5]=<loadbalancerip-5>,cluster.externalAccess.service.loadBalancerIP[6]=,cluster.nodes=7,cluster.init=false oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts. Important here to provide the loadBalancerIP parameters for the new nodes empty to not get an index error.

As we want to add a new node, we are setting cluster.nodes=7 and we leave empty the LoadBalancerIP for the new node, so the cluster will provide the correct one. VALKEY_PASSWORD is the password obtained with the command that appears after the first installation of the Helm Chart. At this point, you will have a new Valkey Pod that will remain in crashLoopBackOff state until we provide the LoadBalancerIP for the new service. Now, wait until the cluster provides the new LoadBalancerIP for the new service and perform the second upgrade:

helm upgrade <release> --set "password=${VALKEY_PASSWORD},cluster.externalAccess.enabled=true,cluster.externalAccess.service.type=LoadBalancer,cluster.externalAccess.service.loadBalancerIP[0]=<loadbalancerip-0>,cluster.externalAccess.service.loadBalancerIP[1]=<loadbalancerip-1>,cluster.externalAccess.service.loadBalancerIP[2]=<loadbalancerip-2>,cluster.externalAccess.service.loadBalancerIP[3]=<loadbalancerip-3>,cluster.externalAccess.service.loadBalancerIP[4]=<loadbalancerip-4>,cluster.externalAccess.service.loadBalancerIP[5]=<loadbalancerip-5>,cluster.externalAccess.service.loadBalancerIP[6]=<loadbalancerip-6>,cluster.nodes=7,cluster.init=false,cluster.update.addNodes=true,cluster.update.newExternalIPs[0]=<load-balancerip-6>" oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts.

Note we are providing the new IPs at cluster.update.newExternalIPs, the flag cluster.update.addNodes=true to enable the creation of the Job that adds a new node and now we are setting the LoadBalancerIP of the new service instead of leave it empty.

NOTE: To avoid the creation of the Job that initializes the Valkey Cluster again, you will need to provide cluster.init=false.

Scale down the cluster

To scale down the Valkey Cluster, follow these steps:

First perform a normal upgrade setting the cluster.nodes value to the desired number of nodes. It should not be less than 6 and the difference between current number of nodes and the desired should be less or equal to cluster.replicas to avoid removing primary node an its replicas at the same time. Also it is needed to provide the password using the password. For example, having more than 6 nodes, to scale down the cluster to 6 nodes:

helm upgrade --timeout 600s <release> --set "password=${VALKEY_PASSWORD},cluster.nodes=6" .

The cluster will continue working during the update as long as the quorum is not lost.

NOTE: To avoid the creation of the Job that initializes the Valkey Cluster again, you will need to provide cluster.init=false.

Once all the nodes are ready, get the list of nodes in the cluster using the CLUSTER NODES command. You will see references to the ones that were removed. Write down the node IDs of the nodes that show fail. In the following example the cluster scaled down from 7 to 6 nodes.

valkey-cli -a $VALKEY_PASSWORD CLUSTER NODES

...
b23bcffa1fd64368d445c1d9bd9aeb92641105f7 10.0.0.70:6379@16379 slave,fail - 1645633139060 0 0 connected
...

In each cluster node, execute the following command. Replace the NODE_ID placeholder.

valkey-cli -a $VALKEY_PASSWORD CLUSTER FORGET NODE_ID

In the previous example the commands would look like this in each cluster node:

valkey-cli -a $VALKEY_PASSWORD CLUSTER FORGET b23bcffa1fd64368d445c1d9bd9aeb92641105f7

Using password file

To use a password file for Valkey you need to create a secret containing the password.

NOTE: It is important that the file with the password must be called valkey-password

And then deploy the Helm Chart using the secret name as parameter:

usePassword=true
usePasswordFile=true
existingSecret=valkey-password-secret
metrics.enabled=true

Securing traffic using TLS

TLS support can be enabled in the chart by specifying the tls. parameters while creating a release. The following parameters should be configured to properly enable the TLS support in the cluster:

• tls.enabled: Enable TLS support. Defaults to false
• tls.existingSecret: Name of the secret that contains the certificates. No defaults.
• tls.certFilename: Certificate filename. No defaults.
• tls.certKeyFilename: Certificate key filename. No defaults.
• tls.certCAFilename: CA Certificate filename. No defaults.

For example:

First, create the secret with the certificates files:

kubectl create secret generic certificates-tls-secret --from-file=./cert.pem --from-file=./cert.key --from-file=./ca.pem

Then, use the following parameters:

tls.enabled="true"
tls.existingSecret="certificates-tls-secret"
tls.certFilename="cert.pem"
tls.certKeyFilename="cert.key"
tls.certCAFilename="ca.pem"

Sidecars and Init Containers

If you have a need for additional containers to run within the same pod as Valkey (e.g. an additional metrics or logging exporter), you can do so via the sidecars config parameter. Simply define your container according to the Kubernetes container spec.

sidecars:
  - name: your-image-name
    image: your-image
    imagePullPolicy: Always
    ports:
      - name: portname
       containerPort: 1234

Similarly, you can add extra init containers using the initContainers parameter.

initContainers:
  - name: your-image-name
    image: your-image
    imagePullPolicy: Always
    ports:
      - name: portname
        containerPort: 1234

Adding extra environment variables

In case you want to add extra environment variables (useful for advanced operations like custom init scripts), you can use the extraEnvVars property.

extraEnvVars:
  - name: VALKEY_WHATEVER
    value: value

Alternatively, you can use a ConfigMap or a Secret with the environment variables. To do so, use the extraEnvVarsCM or the extraEnvVarsSecret values.

Metrics

The chart optionally can start a metrics exporter for prometheus. The metrics endpoint (port 9121) is exposed in the service. Metrics can be scraped from within the cluster using something similar as the described in the example Prometheus scrape configuration. If metrics are to be scraped from outside the cluster, the Kubernetes API proxy can be utilized to access the endpoint.

Host Kernel Settings

Valkey may require some changes in the kernel of the host machine to work as expected, in particular increasing the somaxconn value and disabling transparent huge pages. To do so, you can set up a privileged initContainer with the sysctlImage config values, for example:

sysctlImage:
  enabled: true
  mountHostSys: true
  command:
    - /bin/sh
    - -c
    - |-
      sysctl -w net.core.somaxconn=10000
      echo never > /host-sys/kernel/mm/transparent_hugepage/enabled

Alternatively, for Kubernetes 1.12+ you can set podSecurityContext.sysctls which will configure sysctls for primary and replica pods. Example:

podSecurityContext:
  sysctls:
  - name: net.core.somaxconn
    value: "10000"

Note that this will not disable transparent huge tables.

Helm Upgrade

By default cluster.init will be set to true in order to initialize the Valkey Cluster in the first installation. If for testing purposes you only want to deploy or upgrade the nodes but avoiding the creation of the cluster you can set cluster.init to false.

Backup and restore

To back up and restore Valkey Cluster Helm chart deployments on Kubernetes, you need to back up the persistent volumes from the source deployment and attach them to a new deployment using Velero, a Kubernetes backup/restore tool.

These are the steps you will usually follow to back up and restore your Valkey Cluster database cluster data:

• Install Velero on the source and destination clusters.
• Use Velero to back up the PersistentVolumes (PVs) used by the deployment on the source cluster.
• Use Velero to restore the backed-up PVs on the destination cluster.
• Create a new deployment on the destination cluster with the same chart, deployment name, credentials and other parameters as the original. This new deployment will use the restored PVs and hence the original data.

NetworkPolicy

To enable network policy for Valkey, install a networking plugin that implements the Kubernetes NetworkPolicy spec, and set networkPolicy.enabled to true.

For Kubernetes v1.5 & v1.6, you must also turn on NetworkPolicy by setting the DefaultDeny namespace annotation. Note: this will enforce policy for all pods in the namespace:

kubectl annotate namespace default "net.beta.kubernetes.io/network-policy={\"ingress\":{\"isolation\":\"DefaultDeny\"}}"

With NetworkPolicy enabled, only pods with the generated client label will be able to connect to Valkey. This label will be displayed in the output after a successful install.

With networkPolicy.ingressNSMatchLabels pods from other namespaces can connect to valkey. Set networkPolicy.ingressNSPodMatchLabels to match pod labels in matched namespace. For example, for a namespace labeled valkey=external and pods in that namespace labeled valkey-client=true the fields should be set:

networkPolicy:
  enabled: true
  ingressNSMatchLabels:
    valkey: external
  ingressNSPodMatchLabels:
    valkey-client: true

Setting Pod’s affinity

This chart allows you to set your custom affinity using the XXX.affinity paremeter(s). Find more information about Pod’s affinity in the kubernetes documentation.

As an alternative, you can use of the preset configurations for pod affinity, pod anti-affinity, and node affinity available at the bitnami/common chart. To do so, set the XXX.podAffinityPreset, XXX.podAntiAffinityPreset, or XXX.nodeAffinityPreset parameters.

Persistence

By default, the chart mounts a Persistent Volume at the /bitnami path. The volume is created using dynamic volume provisioning.

If persistence is disabled, an emptyDir volume is used. This is only recommended for testing environments because the required information included in the nodes.conf file is missing. This file contains the relationship between the nodes and the cluster. For example, if any node is down or faulty, when it starts again, it is a self-proclaimed primary and also acts as an independent node outside the main cluster as it doesn’t have the necessary information to connect to it.

To reconnect the failed node, run the following:

See nodes.sh

$ cat /bitnami/valkey/data/nodes.sh
declare -A host_2_ip_array=([valkey-node-0]="192.168.192.6" [valkey-node-1]="192.168.192.2" [valkey-node-2]="192.168.192.4" [valkey-node-3]="192.168.192.5" [valkey-node-4]="192.168.192.3" [valkey-node-5]="192.168.192.7" )

Run valkey-cli and run CLUSTER MEET to any other node in the cluster. Now the node has connected to the main cluster.

$ REDISCLI_AUTH=bitnami valkey-cli
127.0.0.1:6379> cluster meet 192.168.192.7 6379
OK

Parameters

Global parameters

Valkey Cluster Common parameters

Valkey statefulset parameters

Cluster update job parameters

Cluster management parameters

Metrics sidecar parameters

Sysctl Image parameters

Specify each parameter using the --set key=value[,key=value] argument to helm install. For example,

helm install my-release \
  --set password=secretpassword \
    oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts.

The above command sets the Valkey server password to secretpassword.

NOTE: Once this chart is deployed, it is not possible to change the application’s access credentials, such as usernames or passwords, using Helm. To change these application credentials after deployment, delete any persistent volumes (PVs) used by the chart and re-deploy it, or use the application’s built-in administrative tools if available.

Alternatively, a YAML file that specifies the values for the parameters can be provided while installing the chart. For example,

helm install my-release -f values.yaml oci://REGISTRY_NAME/REPOSITORY_NAME/valkey-cluster

Note: You need to substitute the placeholders REGISTRY_NAME and REPOSITORY_NAME with a reference to your Helm chart registry and repository. For example, in the case of Bitnami, you need to use REGISTRY_NAME=registry-1.docker.io and REPOSITORY_NAME=bitnamicharts. Tip: You can use the default values.yaml Note for minikube users: Current versions of minikube (v0.24.1 at the time of writing) provision hostPath persistent volumes that are only writable by root. Using chart defaults cause pod failure for the Valkey pod as it attempts to write to the /bitnami directory. See minikube issue 1990 for more information.

Troubleshooting

Find more information about how to deal with common errors related to Bitnami’s Helm charts in this troubleshooting guide.

Upgrading

To 2.1.0

This version introduces image verification for security purposes. To disable it, set global.security.allowInsecureImages to true. More details at GitHub issue.

To 2.0.0

This major updates all the references from master/slave to primary/replica to follow the upstream project strategy:

• The term master has been replaced by the term primary. Therefore, parameters prefixed with master are now prefixed with primary.
• Environment variables previously prefixed as VALKEY_MASTER or VALKEY_SENTINEL_MASTER use VALKEY_PRIMARY and VALKEY_SENTINEL_PRIMARY now.

Consequences:

Backwards compatibility is not guaranteed. To upgrade to 2.0.0, install a new release of the Valkey chart, and migrate the data from your previous release. You have to create a backup of the database, and restore it on the new release as explained in the Backup and restore section.

License

Copyright © 2025 Broadcom. The term “Broadcom” refers to Broadcom Inc. and/or its subsidiaries.

Licensed under the Apache License, Version 2.0 (the “License”); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an “AS IS” BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.