The statsPort is not correctly set for HostNetwork endpointPublishingStrategyWhen we change the httpPort from 80 to 85 and statsPort from 1936 to 1939 on the default router like here: # oc get IngressController default -n openshift-ingress-operator
...
 clientTLS:
    clientCA:
      name: ""
    clientCertificatePolicy: ""
  endpointPublishingStrategy:
    hostNetwork:
      httpPort: 85
      httpsPort: 443
      statsPort: 1939
    type: HostNetwork
...
status:
...  
endpointPublishingStrategy:
    hostNetwork:
      httpPort: 85
      httpsPort: 443
      protocol: TCP
      statsPort: 1939
 
We can see that the route pods get restarted:# oc get pod -n openshift-ingress
NAME                              READY   STATUS    RESTARTS   AGE
router-default-5b96855754-2wnrp   1/1     Running   0          1m
router-default-5b96855754-9c724   1/1     Running   0          2mThe pods are configured correctly:# oc get pod router-default-5b96855754-2wnrp -o yaml
...
spec:
  containers:
  - env:
    - name: ROUTER_SERVICE_HTTPS_PORT
      value: "443"
    - name: ROUTER_SERVICE_HTTP_PORT
      value: "85"
    - name: STATS_PORT
      value: "1939"
...
    livenessProbe:
      failureThreshold: 3
      httpGet:
        host: localhost
        path: /healthz
        port: 1939
        scheme: HTTP
...
    ports:
    - containerPort: 85
      hostPort: 85
      name: http
      protocol: TCP
    - containerPort: 443
      hostPort: 443
      name: https
      protocol: TCP
    - containerPort: 1939
      hostPort: 1939
      name: metrics
      protocol: TCPBut the endpoint is incorrect:# oc get ep router-internal-default -o yaml
...
apiVersion: v1
items:
- apiVersion: v1
  kind: Endpoints
  metadata:
    creationTimestamp: "2022-12-02T13:34:48Z"
    labels:
      ingresscontroller.operator.openshift.io/owning-ingresscontroller: default
    name: router-internal-default
    namespace: openshift-ingress
    resourceVersion: "23216275"
    uid: 50c00fc0-08e5-4a6a-a7eb-7501fa1a7ba6
  subsets:
  - addresses:
    - ip: 10.74.211.203
      nodeName: worker-0.rhodain01.lab.psi.pnq2.redhat.com
      targetRef:
        kind: Pod
        name: router-default-5b96855754-2wnrp
        namespace: openshift-ingress
        uid: eda945b9-9061-4361-b11a-9d895fee0003
    - ip: 10.74.211.216
      nodeName: worker-1.rhodain01.lab.psi.pnq2.redhat.com
      targetRef:
        kind: Pod
        name: router-default-5b96855754-9c724
        namespace: openshift-ingress
        uid: 97a04c3e-ddea-43b7-ac70-673279057929
    ports:
    - name: metrics
      port: 1936
      protocol: TCP
    - name: https
      port: 443
      protocol: TCP
    - name: http
      port: 85
      protocol: TCP
kind: List
metadata:
  resourceVersion: ""
  selfLink: ""Notice that the https port is correctly set to 85, but the stats port is still set to 1936 and not to 1939. That is a problem as the metrics target endpoint is reported as down with an error message:    Get "https://10.74.211.203:1936/metrics": dial tcp 10.74.211.203:1936: connect: connection refusedWhen the EP is corrected and the ports are changed to:
  ports:
  - name: metrics
    port: 1939
    protocol: TCPthe metrics target endpoint is picked up correctly and the metrics are scribed works as expected

100%

Set endpointPublishingStrategy and modify the nodeport for statPort:

endpointPublishingStrategy:
    hostNetwork:
      httpPort: 85
      httpsPort: 443
      protocol: TCP
      statsPort: 1939

Stats are scribed from the standard port and not the one specified.

The endpoint object is pointing to the specified port.