Kubernetes is the recommended way to run ParadeDB in production. Both ParadeDB Community and Enterprise binaries can be deployed on Kubernetes. This guide uses the ParadeDB Helm Chart. The chart is also available on Artifact Hub.

​
Prerequisites

This guide assumes you have installed Helm and have a Kubernetes cluster running v1.25+. For local testing, we recommend Minikube.

​
Install the Prometheus Stack

The ParadeDB Helm chart supports monitoring via Prometheus and Grafana. To enable this, you need to have the Prometheus CRDs installed before installing the CloudNativePG operator. If you do not yet have the Prometheus CRDs installed on your Kubernetes cluster, you can install it with: 
helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm upgrade --atomic --install prometheus-community \
--create-namespace \
--namespace prometheus-community \
--values https://raw.githubusercontent.com/cloudnative-pg/cloudnative-pg/main/docs/src/samples/monitoring/kube-stack-config.yaml \
prometheus-community/kube-prometheus-stack

​
Install the CloudNativePG Operator

Skip this step if the CloudNativePG operator is already installed in your cluster. If you do not wish to monitor your cluster, omit the --set commands. 
helm repo add cnpg https://cloudnative-pg.github.io/charts
helm upgrade --atomic --install cnpg \
--create-namespace \
--namespace cnpg-system \
--set monitoring.podMonitorEnabled=true \
--set monitoring.grafanaDashboard.create=true \
cnpg/cloudnative-pg

​
Start a ParadeDB CNPG Cluster

Create a values.yaml and configure it to your requirements. Here is a basic example: 
type: paradedb
mode: standalone

cluster:
  instances: 1
  storage:
    size: 256Mi
If you are using ParadeDB Enterprise, instances should be set to a number greater than 1 for high availability.
Next, create a namespace for this step or use an existing namespace. The namespace can be any value. 
kubectl create namespace <your-namespace>
For ParadeDB Enterprise, you should have received an enterprise Docker username and personal access token. The following step passes these credentials to Kubernetes and should be skipped if you are deploying ParadeDB Community.
ParadeDB Enterprise
kubectl create secret docker-registry paradedb-enterprise-registry-cred
--namespace <your-namespace>
--docker-server="https://index.docker.io/v1/"
--docker-username="<enterprise_docker_username>"
--docker-password="<enterprise_docker_access_token>"
Finally, launch the ParadeDB cluster. 
helm repo add paradedb https://paradedb.github.io/charts
helm upgrade --atomic --install paradedb \
--namespace <your-namespace> \
--values values.yaml \
--set cluster.monitoring.enabled=true \
paradedb/paradedb

​
Connect to the Cluster

The command to connect to the primary instance of the cluster will be printed in your terminal. If you do not modify any settings, it will be: 
kubectl --namespace paradedb exec --stdin --tty services/paradedb-rw -- bash
This will launch a Bash shell inside the instance. You can connect to the ParadeDB database via psql with: 
psql -d paradedb

​
Connect to the Grafana Dashboard

To connect to the Grafana dashboard for your cluster, we suggested port forwarding the Kubernetes service running Grafana to localhost: 
kubectl --namespace prometheus-community port-forward svc/prometheus-community-grafana 3000:80
You can then access the Grafana dasbhoard at http://localhost:3000/ using the credentials admin as username and prom-operator as password. These default credentials are defined in the kube-stack-config.yaml file used as the values.yaml file in Installing the Prometheus CRDs and can be modified by providing your own values.yaml file. A more detailed guide on monitoring the cluster can be found in the CloudNativePG documentation.