Ingress GCE is not working due to health check being failed for backend service

[yugandharreddy1]

We see than when we install with NIGINX everything is working fine but then changing to ingress GCE with below annotations its creating 2 services one is healthy and other is unhealthy.

ingress:
  annotations:
    - kubernetes.io/ingress.class: gce
    - kubernetes.io/ingress.allow-http: "true"
    - gcp.ingress.kubernetes.io/target-type: ip

below are the services which are being created

by default the health check is checking on HTTP 80 and when i change it to TCP 3128 its becoming healthy and i was able to access the ingress URL.

i dont know from which service readiness probe i am getting the healthcheck created as below i didnt mention anywhere HTTP 80

Squid-svc.yaml

kind: Service
apiVersion: v1
metadata:
  name: {{ .Chart.Name }}-service
  namespace: {{ .Values.namespace}}
  labels:
spec:
  type: ClusterIP
  selector:
    app: {{ .Release.Name }}-{{ .Chart.Name }}
  ports:
    - protocol: TCP
      port: {{ .Values.svc.port }}
      targetPort: {{.Values.svc.targetPort}}
      name: http
      nodePort: null

app-ingress.yaml

{{- if .Values.ingress.enabled -}}
{{- if .Capabilities.APIVersions.Has "networking.k8s.io/v1/Ingress" }}
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: app-ingress
  labels:
{{ include "app.labels" . | indent 4 }}
  annotations:
    {{- range $annotation := .Values.ingress.annotations }}
    {{- range $key, $value := $annotation }}
    {{ $key }}: {{ $value | quote }}
    {{- end }}
    {{- end }}
spec:
{{- if .Values.ingress.tls.enabled }}
  tls:
    - hosts:
        - {{ .Values.ingress.host }}
      secretName: {{ .Values.ingress.tls.secretName | default "app-tls-secrets" }}
{{- end }}
  rules:
  - host: {{ .Values.ingress.host }}
    http:
      paths:
      - pathType: ImplementationSpecific 
        backend:
          service:
            {{- if and (.Values.squid) (.Values.squid.enabled) }}
            name: {{ .Values.squid.serviceName | default "squid-service" }}
            port: 
              number: {{ .Values.squid.port | default 3128 }}
            {{- else}}
            name: app-ext
            port: 
              number: {{ .Values.app.port | default 8080 }}
            {{- end }}
            
{{- else if .Capabilities.APIVersions.Has "extensions/v1beta1/Ingress" }}
apiVersion: extensions/v1beta1
kind: Ingress
metadata:
  name: app-ingress
  labels:
{{ include "app.labels" . | indent 4 }}
  annotations:
    {{- range $annotation := .Values.ingress.annotations }}
    {{- range $key, $value := $annotation }}
    {{ $key }}: {{ $value | quote }}
    {{- end }}
    {{- end }}
spec:
{{- if .Values.ingress.tls.enabled }}
  tls:
    - hosts:
        - {{ .Values.ingress.host }}
      secretName: {{ .Values.ingress.tls.secretName | default "app-tls-secrets" }}
{{- end }}
  rules:
  - host: {{ .Values.ingress.host }}
    http:
      paths:
      - backend:
        {{- if and (.Values.squid) (.Values.squid.enabled) }}
          serviceName: {{ .Values.squid.serviceName | default "squid-service" }}
          servicePort: {{ .Values.squid.port | default 3128 }}
        {{- else}}
          serviceName: app-ext
          servicePort: {{ .Values.app.port | default 8080 }}
        {{- end }}
{{- end }}
{{- end }}

squid.yaml

{{- $appLabelValue := printf "%s-%s" .Release.Name .Chart.Name -}}
{{- $engineTimeoutMillis := default 12000 .Values.engineTimeoutMillis -}}
{{- $terminationGracePeriodSeconds := div (add $engineTimeoutMillis 2000 ) 1000 -}}
{{- $containerName := "squid" -}}

apiVersion: apps/v1
kind: Deployment
metadata:
  name: {{ $containerName }}
  namespace: {{ .Values.namespace}}
spec:
  replicas: {{ .Values.replicaCount }}
  strategy:
    {{- include "deployment.strategy" . | nindent 4 }}
  selector:
    matchLabels:
      app: {{ $appLabelValue }}
  template:
    metadata:
      labels:
        app: {{ $appLabelValue }}
      annotations:
        checksum/config: {{ include (print $.Chart.Name "/templates/squid-config.yaml") . | sha256sum }}
    spec:
      {{- if .Values.image.pullSecret }}
      imagePullSecrets:
        - name: {{ .Values.image.pullSecret }}
      {{- end }}
      terminationGracePeriodSeconds: {{ $terminationGracePeriodSeconds }}
      containers:
        - name: {{ $containerName }}
          image: {{ template "squid.image" .Values }}
          imagePullPolicy: {{ .Values.image.pullPolicy }}
          securityContext:
            {{- include "getDefaultSecurityContextProperties" . | nindent 12 }}
{{- if eq .Values.resources.enabled true }}
          resources:
            {{ .Values.resources | toYaml | indent 12 | trim }}
{{- end }}
          terminationMessagePolicy: FallbackToLogsOnError
          ports:
            - containerPort: 3128
          volumeMounts:
            - mountPath: /var/opt/{{ .Chart.Name }}
              name: {{ .Chart.Name }}-volume
          readinessProbe:
            tcpSocket:              
              port: 3128
            initialDelaySeconds: 30
{{- if not .Values.debugger }}
          livenessProbe:
            tcpSocket:
              port: 3128
            initialDelaySeconds: 30
{{- end }}
      volumes:
        - name: {{ .Chart.Name }}-volume
          configMap:
            name: {{.Release.Name}}-{{ .Chart.Name }}-config

app-service.yml

apiVersion: v1
kind: Service
metadata:
  name: {{ .Values.container.name }}
  labels:
{{ include "app.labels" . | indent 4 }}
spec:
  # This is needed so that DNS resolves between containers in a statefulset 
  # before they are marked "Ready", as containers are only marked "Ready" after
  # they are running.  Our containers need to talk to each other during start-up.
  publishNotReadyAddresses: true
  clusterIP: None
  ports:
    - name: http
      port: {{ .Values.app.port | default 8080 }}
      targetPort: 8080
  {{- if .Values.app.portRange }}
  {{- $ports := split "-" .Values.app.portRange -}}
  {{- range untilStep ($ports._0|int) ($ports._1|int) 1 }}
    - name: tcp-{{ . }}
      port: {{ . }}
      protocol: TCP
      targetPort: {{ . }}
  {{- end }}
  {{- end }}
  selector:
    app: {{ template "app.name" . }}

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale