pulsar-and-celte

Deploying Apache Pulsar for Celte (Helm + Kubernetes)

This guide shows how to deploy a production-ready Apache Pulsar cluster for Celte using Helm with the following values. It enables all core components and sets resource requests/limits and storage.

Prerequisites

• A Kubernetes cluster (3+ worker nodes recommended)
• kubectl and Helm installed
• A default StorageClass or local-path provisioner installed (for storage.className: local-path)
• Cluster nodes reachable on the advertised broker IP if you set a public advertisedAddress

Namespace and chart repo

kubectl create namespace pulsar
helm repo add apache https://pulsar.apache.org/charts
helm repo update

Save values file

Create pulsar-values.yaml with the content below.

# Enable core components
components:
  zookeeper: true
  bookkeeper: true
  broker: true
  proxy: true
  functions: true
  pulsar_manager: true

# Anti-affinity helper template
global:
  podAntiAffinity:
    enabled: true
    topologyKey: kubernetes.io/hostname  # spread across VMs

zookeeper:
  replicaCount: 3
  resources:
    requests:
      cpu: "200m"
      memory: "1Gi"
    limits:
      cpu: "500m"
      memory: "2Gi"
  persistence:
    enabled: true
    size: 10Gi
  affinity:
    podAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - labelSelector:
          matchLabels:
            app.kubernetes.io/name: pulsar-zookeeper
        topologyKey: kubernetes.io/hostname

bookkeeper:
  replicaCount: 3
  resources:
    requests:
      cpu: "1"
      memory: "2Gi"
    limits:
      cpu: "2"
      memory: "3Gi"
  persistence:
    enabled: true
    size: 30Gi
  affinity:
    podAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - labelSelector:
          matchLabels:
            app.kubernetes.io/name: pulsar-bookkeeper
        topologyKey: kubernetes.io/hostname

broker:
  replicaCount: 3
  resources:
    requests:
      cpu: "1"
      memory: "2Gi"
    limits:
      cpu: "2"
      memory: "3Gi"
  configuration:
    advertisedAddress: "57.128.60.39"
    brokerServicePort: 6650
    brokerServicePortTls: 6651
  affinity:
    podAntiAffinity:
      requiredDuringSchedulingIgnoredDuringExecution:
      - labelSelector:
          matchLabels:
            app.kubernetes.io/name: pulsar-broker
        topologyKey: kubernetes.io/hostname

proxy:
  replicaCount: 1
  service:
    type: NodePort
    ports:
      http: 8080
    configuration:
      brokerServiceURL: pulsar://pulsar-proxy.pulsar.svc.cluster.local:6650
  resources:
    requests:
      cpu: "500m"
      memory: "1Gi"
    limits:
      cpu: "1"
      memory: "2Gi"

functions:
  replicaCount: 1
  resources:
    requests:
      cpu: "500m"
      memory: "1Gi"
    limits:
      cpu: "1"
      memory: "2Gi"

pulsar_manager:
  enabled: true
  resources:
    requests:
      cpu: "200m"
      memory: "512Mi"
    limits:
      cpu: "500m"
      memory: "1Gi"

# Storage backend
storage:
  className: local-path

Install

helm install pulsar apache/pulsar \
  --namespace pulsar \
  -f pulsar-values.yaml

Verify

kubectl -n pulsar get pods
kubectl -n pulsar get svc

• Brokers should be 3/3 ready, Zookeeper 3/3, BookKeeper 3/3.
• The proxy service is NodePort exposing HTTP 8080 and broker service URL internally.

Notes for Celte

• Clients connect to the broker via the proxy or directly to brokers on port 6650. Ensure firewall/NAT allows access to the advertisedAddress (57.128.60.39) if clients are outside the cluster.
• Storage class local-path requires a local-path-provisioner. Replace storage.className with your cloud provider’s storage class in managed clusters.
• For TLS or external load balancers, set appropriate brokerServicePortTls and configure proxy/broker ingress accordingly.