HPA created pod on emqx-replicant deployment is terminated

[johnlim2019]

hey I am trying to configure the emqx-replicant deployment with k8s hpa, but operator seems to override the hpa's desired number of replicas. So when hpa creates a new pod, operator seems to terminate that pod so the number of replicas is stuck with the initial number from setup.

Is there some configuration i am missing out? any help is welcome?

i am using the open source emqx version 5.0.26 and the api version v2alpha1.

The emqx operator is deployed on a local kind cluster, k8s v1.24.13

I installed it following the guide

helm upgrade --install cert-manager jetstack/cert-manager \
  --namespace cert-manager \
  --create-namespace \
  --set installCRDs=true 
helm upgrade --install emqx-operator emqx/emqx-operator \
  --namespace emqx-operator-system \
  --create-namespace 
kubectl wait --for=condition=Ready pods -l "control-plane=controller-manager" -n emqx-operator-system
kubectl apply -f ./emqx-operator.yaml

apiVersion: apps.emqx.io/v2alpha1
kind: EMQX
metadata:
  name: emqx
spec:
  image: emqx:5.0.26
  bootstrapConfig: | 
    mqtt {
      max_packet_size = "256MB",
    },
    session {
      max_mqueue_len = 10000
    },
    log {
     console_handler {
      level = error
      }
    }
    
  coreTemplate:
    spec:
      replicas: 1
      ports:
        # prometheus monitor requires the pod must name the target port 
        - containerPort: 18083
          name: dashboard
      resources:
        requests:
          cpu: "0.1"
          memory: 141Mi
        limits:
          cpu: "0.25"
          memory: "256Mi"
  replicantTemplate:
    spec:
      replicas: 3
      ports:
        # prometheus monitor requires the pod must name the target port
        - containerPort: 18083
          name: dashboard
      resources:
        requests:
          cpu: "0.25"
          memory: 512Mi
        limits:
          cpu: "1"
          memory: "744Mi"
  listenersServiceTemplate:
    spec:
      type: LoadBalancer
      ports:
      - name: mqtt
        nodePort: 30000
        port: 1883
        targetPort: 1883
      - name: mqttssl
        nodePort: 30001
        port: 8883
        targetPort: 8883
  dashboardServiceTemplate:
    spec:
      type: LoadBalancer
      ports:  
        - name: dashboard
          nodePort: 30002
          port: 18083
          targetPort: dashboard

[Rory-Z]

Yes, we do not currently support HPA, because scaling down the replicant node will cause the MQTT client to reconnect.