Exposing Kubernetes Service and LoadBalancer on GCP

Overview

This project provides sample code to understand the differences between Ingress/ClusterIP, LoadBalancer, and NodePort on GKE. To evenly distribute traffic to Pods, it's recommended to use the container-native load balancer and Ingress by GCP.

ingress-neg-api-template.yaml
loadbalancer-type-api.yaml
nodeport-type-api-template.yaml

kubectl get svc

NAME                    TYPE           CLUSTER-IP      EXTERNAL-IP      PORT(S)        AGE
ingress-neg-api         ClusterIP      10.25.130.238   <none>           8000/TCP       40m
loadbalancer-type-api   LoadBalancer   10.25.131.128   34.133.110.139   80:31000/TCP   31m
nodeport-type-api       NodePort       10.25.129.2     <none>           80:32000/TCP   17m

	Ingress/NEG	LoadBalancer	NodePort
Type of K8s Service	ClusterIP	LoadBalancer	NodePort
Type of Load Balancer	HTTP(S) load balancer	Network load balancer(TCP/UDP)	X
Use a NodePort	X	O	O
Endpoint	Frontend external IP of load balancer	Service external IP	Node external IP

The Network Endpoint Group (NEG) is a configuration object that specifies a group of backend endpoints or services.

Objectives

Learn about the following topics:

Differences among Ingress, LoadBalancer, and NodePort on GKE
Manifests for Deployment, Service, Ingress, BackendConfig, and HorizontalPodAutoscaler
How to use the container-native load balancer with a manifest

Step1: Create a GKE cluster and namespaces
Step2: Build and push to GCR
Step3: Ingress with Network Endpoint Group (NEG)
- Manifest
- Deploy ingress-neg-api
- Screenshots
Step4: LoadBalancer Type with NodePort
- Manifest
- Deploy loadbalancer-type-api
- Screenshots
Step5: NodePort Type
- Manifest
- Deploy nodeport-type-api
- Create a firewall rule for the node port
Cleanup

Prerequisites

Installation

Install the gcloud CLI
Install kubectl and configure cluster access
Install the jq

Set environment variables

COMPUTE_ZONE="us-central1"
# replace with your project
PROJECT_ID="sample-project"

Set GCP project

gcloud config set project ${PROJECT_ID}
gcloud config set compute/zone ${COMPUTE_ZONE}

1. Create a GKE cluster

Create an Autopilot GKE cluster. It may take around 9 minutes.

gcloud container clusters create-auto sample-cluster --region=${COMPUTE_ZONE}
gcloud container clusters get-credentials sample-cluster

2. Build and push to GCR

cd ./app
docker build -t python-ping-api . --platform linux/amd64
docker tag python-ping-api:latest gcr.io/${PROJECT_ID}/python-ping-api:latest

gcloud auth configure-docker
docker push gcr.io/${PROJECT_ID}/python-ping-api:latest

3. Ingress with Network Endpoint Group(NEG)

3.1 Manifest

ingress-neg-api-template.yaml:

Kind	Element	Value
Service	spec.type	ClusterIP
Service	metadata.annotations	cloud.google.com/neg: '{"ingress": true}'
Ingress	metadata.annotations	kubernetes.io/ingress.class: gce

The container-native load balancing is default from GKE cluster v1.17+ and does not require an explicit cloud.google.com/neg: '{"ingress": true}' Service annotation.

apiVersion: v1
kind: Service
metadata:
  name: ingress-neg-api
  annotations:
    app: ingress-neg-api
    cloud.google.com/neg: '{"ingress": true}'
    cloud.google.com/backend-config: '{"default": "ingress-neg-api-backend-config"}'
spec:
  selector:
    app: ingress-neg-api
  type: ClusterIP
  ports:
    - port: 8000
      targetPort: 8000
      protocol: TCP
---
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: ingress-neg-api-ingress
  annotations:
    app: ingress-neg-api
    kubernetes.io/ingress.class: gce
spec:
  rules:
  - http:
        paths:
          - path: /*
            pathType: ImplementationSpecific
            backend:
              service:
                name: ingress-neg-api
                port:
                  number: 8000
---
apiVersion: cloud.google.com/v1
kind: BackendConfig
metadata:
  name: ingress-neg-api-backend-config
spec:
  healthCheck:
    checkIntervalSec: 10
    timeoutSec: 10
    healthyThreshold: 2
    unhealthyThreshold: 5
    port: 8000
    type: HTTP
    requestPath: /ping

3.2 Deploy ingress-neg-api

Create and deploy a K8s Deployment, Service, Ingress, GKE BackendConfig, and HorizontalPodAutoscaler using the template files. It may take around 5 minutes to create a load balancer, including health checks.

sed -e "s|<project-id>|${PROJECT_ID}|g" ingress-neg-api-template.yaml > ingress-neg-api.yaml
cat ingress-neg-api.yaml

kubectl apply -f ingress-neg-api.yaml

Confirm Pod logs and service configuration after deployment:

kubectl logs -l app=ingress-neg-api

kubectl get service ingress-neg-api --output yaml

apiVersion: v1
kind: Service
metadata:
  annotations:
    app: ingress-neg-api
    cloud.google.com/backend-config: '{"default": "ingress-neg-api-backend-config"}'
    cloud.google.com/neg: '{"ingress": true}'
    cloud.google.com/neg-status: '{"network_endpoint_groups":{"8000":"k8s1-d0ddc61c-default-ingress-neg-api-8000-276c73a2"},"zones":["us-central1-b","us-central1-c"]}'
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"Service","metadata":{"annotations":{"app":"ingress-neg-api","cloud.google.com/backend-config":"{\"default\": \"ingress-neg-api-backend-config\"}","cloud.google.com/neg":"{\"ingress\": true}"},"name":"ingress-neg-api","namespace":"default"},"spec":{"ports":[{"port":8000,"protocol":"TCP","targetPort":8000}],"selector":{"app":"ingress-neg-api"},"type":"ClusterIP"}}
  creationTimestamp: "2022-11-24T04:01:41Z"
  name: ingress-neg-api
  namespace: default
  resourceVersion: "75752"
  uid: 5687946c-096f-4c77-b945-4011e86221a0
spec:
  clusterIP: 10.25.130.238
  clusterIPs:
  - 10.25.130.238
  internalTrafficPolicy: Cluster
  ipFamilies:
  - IPv4
  ipFamilyPolicy: SingleStack
  ports:
  - port: 8000
    protocol: TCP
    targetPort: 8000
  selector:
    app: ingress-neg-api
  sessionAffinity: None
  type: ClusterIP
status:
  loadBalancer: {}

kubectl get svc ingress-neg-api -o=jsonpath="{.metadata.annotations.cloud\.google\.com/neg-status}" | jq

{
  "network_endpoint_groups": {
    "8000": "k8s1-d0ddc61c-default-ingress-neg-api-8000-276c73a2"
  },
  "zones": [
    "us-central1-b",
    "us-central1-c"
  ]
}

3.3 Screenshots

Services & Ingress > SERVICES

https://console.cloud.google.com/kubernetes/discovery
Services & Ingress > Service details > OVERVIEW
Services & Ingress > Service details > DETAILS
Services & Ingress > INGRESS > DETAILS

https://console.cloud.google.com/kubernetes/ingresses
Network services > Load balancing > LOAD BALANCERS

https://console.cloud.google.com/net-services/loadbalancing/list/loadBalancers
Network services > Load balancing > BACKENDS

https://console.cloud.google.com/net-services/loadbalancing/list/backends

4. LoadBalancer Type with NodePort

4.1 Manifest

NOTE: Ingress is not required when creating a Service with the LoadBalancer type.

loadbalancer-type-api-template.yaml:

Kind	Element	Value	Description
Service	spec.type	LoadBalancer
Service	spec.ports.port	80	You can access with `<external-endpoint-ip>:<port>`.
Service	spec.ports.nodePort	31000	This element is optional and GKE will assign a port in 30000-32768 range automatically if you do not aggign it.

apiVersion: v1
kind: Service
metadata:
  name: loadbalancer-type-api
  annotations:
    app: loadbalancer-type-api
spec:
  selector:
    app: loadbalancer-type-api
  type: LoadBalancer
  externalTrafficPolicy: Local
  ports:
    - port: 80
      targetPort: 8000
      nodePort: 31000
      protocol: TCP

4.2 Deploy loadbalancer-type-api

sed -e "s|<project-id>|${PROJECT_ID}|g" loadbalancer-type-api-template.yaml > loadbalancer-type-api.yaml
cat loadbalancer-type-api.yaml

kubectl apply -f loadbalancer-type-api.yaml

Confirm Pod logs and service configuration after deployment:

kubectl logs -l app=loadbalancer-type-api

kubectl get service loadbalancer-type-api --output yaml

GKE create a TCP load balancer and assign an external IP:

status:
  loadBalancer:
    ingress:
    - ip: 34.133.110.139

apiVersion: v1
kind: Service
metadata:
  annotations:
    app: loadbalancer-type-api
    cloud.google.com/neg: '{"ingress":true}'
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"Service","metadata":{"annotations":{"app":"loadbalancer-type-api"},"name":"loadbalancer-type-api","namespace":"default"},"spec":{"externalTrafficPolicy":"Local","ports":[{"nodePort":31000,"port":80,"protocol":"TCP","targetPort":8000}],"selector":{"app":"loadbalancer-type-api"},"type":"LoadBalancer"}}
  creationTimestamp: "2022-11-24T04:10:30Z"
  finalizers:
  - service.kubernetes.io/load-balancer-cleanup
  name: loadbalancer-type-api
  namespace: default
  resourceVersion: "81655"
  uid: e2573c24-fc45-4bb7-9c89-2521a4bf5139
spec:
  allocateLoadBalancerNodePorts: true
  clusterIP: 10.25.131.128
  clusterIPs:
  - 10.25.131.128
  externalTrafficPolicy: Local
  healthCheckNodePort: 31456
  internalTrafficPolicy: Cluster
  ipFamilies:
  - IPv4
  ipFamilyPolicy: SingleStack
  ports:
  - nodePort: 31000
    port: 80
    protocol: TCP
    targetPort: 8000
  selector:
    app: loadbalancer-type-api
  sessionAffinity: None
  type: LoadBalancer
status:
  loadBalancer:
    ingress:
    - ip: 34.133.110.139

4.3 Screenshots

Services & Ingress > SERVICES

https://console.cloud.google.com/kubernetes/discovery
Services & Ingress > Service details > OVERVIEW
Services & Ingress > Service details > DETAILS

5. NodePort Type

5.1 Manifest

nodeport-type-api-template.yaml:

apiVersion: v1
kind: Service
metadata:
  name: nodeport-type-api
  annotations:
    app: nodeport-type-api
spec:
  selector:
    app: nodeport-type-api
  type: NodePort
  ports:
    - port: 80
      targetPort: 8000
      nodePort: 32000
      protocol: TCP

5.2 Deploy nodeport-type-api

sed -e "s|<project-id>|${PROJECT_ID}|g" nodeport-type-api-template.yaml > nodeport-type-api.yaml
cat nodeport-type-api.yaml

kubectl apply -f nodeport-type-api.yaml

Confirm Pod logs and service configuration after deployment:

kubectl logs -l app=nodeport-type-api

kubectl get service nodeport-type-api --output yaml

apiVersion: v1
kind: Service
metadata:
  annotations:
    app: nodeport-type-api
    cloud.google.com/neg: '{"ingress":true}'
    kubectl.kubernetes.io/last-applied-configuration: |
      {"apiVersion":"v1","kind":"Service","metadata":{"annotations":{"app":"nodeport-type-api"},"name":"nodeport-type-api","namespace":"default"},"spec":{"ports":[{"port":80,"protocol":"TCP","targetPort":8000}],"selector":{"app":"nodeport-type-api"},"type":"NodePort"}}
  creationTimestamp: "2022-11-24T04:24:47Z"
  name: nodeport-type-api
  namespace: default
  resourceVersion: "90492"
  uid: 33f1feab-0846-4326-8fc0-e0704194653b
spec:
  clusterIP: 10.25.129.2
  clusterIPs:
  - 10.25.129.2
  externalTrafficPolicy: Cluster
  internalTrafficPolicy: Cluster
  ipFamilies:
  - IPv4
  ipFamilyPolicy: SingleStack
  ports:
  - nodePort: 32000
    port: 80
    protocol: TCP
    targetPort: 8000
  selector:
    app: nodeport-type-api
  sessionAffinity: None
  type: NodePort
status:
  loadBalancer: {}

5.3 Create a firewall rule for node port

Create a firewall rule to allow TCP traffic on your node port:

gcloud compute firewall-rules create test-node-port --allow tcp:32000
gcloud compute firewall-rules describe test-node-port

allowed:
- IPProtocol: tcp
  ports:
  - '32000'
creationTimestamp: '2022-11-23T21:45:10.124-08:00'
description: ''
direction: INGRESS
disabled: false
id: '1317487238481885705'
kind: compute#firewall
logConfig:
  enable: false
name: test-node-port
network: https://www.googleapis.com/compute/v1/projects/moloco-sre/global/networks/default
priority: 1000
selfLink: https://www.googleapis.com/compute/v1/projects/moloco-sre/global/firewalls/test-node-port
sourceRanges:
- 0.0.0.0/0

kubectl get nodes -o custom-columns='NAME:metadata.name,ZONE:metadata.labels.failure-domain\.beta\.kubernetes\.io/zone,CPU:.status.capacity.cpu,MEM:.status.capacity.memory,EXTERNAL-IP:.status.addresses[?(@.type=="ExternalIP")].address'

# or kubectl get nodes --output wide

NAME                                            ZONE            CPU   MEM         EXTERNAL-IP
gk3-sample-cluster-default-pool-21a76107-42lv   us-central1-b   2     4026000Ki   34.71.157.59
gk3-sample-cluster-default-pool-21a76107-mgfr   us-central1-b   2     4026000Ki   34.172.14.138
gk3-sample-cluster-default-pool-7a0947da-blv6   us-central1-c   2     4026000Ki   34.170.106.6
gk3-sample-cluster-default-pool-7a0947da-nkc8   us-central1-c   2     4026008Ki   34.173.168.212

Connecting an external IP with a node port:

EXTERNAL_IP=$(kubectl get nodes -o custom-columns='EXTERNAL-IP:status.addresses[?(@.type=="ExternalIP")].address' | sort | head -n 1)
echo $EXTERNAL_IP
curl http://$EXTERNAL_IP:32000 | jq

{
  "host":"34.170.106.6:32000",
  "message":"ping-api",
  "method":"GET",
  "url":"http://34.170.106.6:32000/"
}

5.4 Screenshots

6. Cleanup

kubectl delete -f app/ingress-neg-api.yaml
kubectl delete -f app/loadbalancer-type-api.yaml
kubectl delete -f app/nodeport-type-api.yaml

gcloud compute firewall-rules delete test-node-port

gcloud container clusters delete sample-cluster

References

Cloud Load Balancing > Documentation > Guides
- Serverless network endpoint groups overview

Google Kubernetes Engine (GKE) > Documentation > Guides

Kubernetes Documentation / Reference / Command line tool (kubectl) / kubectl Cheat Sheet

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

README.md

README.md

Exposing Kubernetes Service and LoadBalancer on GCP

Overview

Objectives

Table of Contents

Prerequisites

Installation

Set environment variables

Set GCP project

1. Create a GKE cluster

2. Build and push to GCR

3. Ingress with Network Endpoint Group(NEG)

3.1 Manifest

3.2 Deploy ingress-neg-api

3.3 Screenshots

4. LoadBalancer Type with NodePort

4.1 Manifest

4.2 Deploy loadbalancer-type-api

4.3 Screenshots

5. NodePort Type

5.1 Manifest

5.2 Deploy nodeport-type-api

5.3 Create a firewall rule for node port

5.4 Screenshots

6. Cleanup

References

Files

README.md

Latest commit

History

README.md

File metadata and controls

Exposing Kubernetes Service and LoadBalancer on GCP

Overview

Objectives

Table of Contents

Prerequisites

Installation

Set environment variables

Set GCP project

1. Create a GKE cluster

2. Build and push to GCR

3. Ingress with Network Endpoint Group(NEG)

3.1 Manifest

3.2 Deploy ingress-neg-api

3.3 Screenshots

4. LoadBalancer Type with NodePort

4.1 Manifest

4.2 Deploy loadbalancer-type-api

4.3 Screenshots

5. NodePort Type

5.1 Manifest

5.2 Deploy nodeport-type-api

5.3 Create a firewall rule for node port

5.4 Screenshots

6. Cleanup

References