Full Stack Application on Google Kubernetes Engine (GKE)

This repository contains the code for a full-stack application deployed on Google Kubernetes Engine (GKE). The application consists of the following services:

Blogs App - GKE Microservices Demo

Made with ❤️ by Sounish Nath using Angular, Golang, Postgresql, Google Kubernetes Engine 🚀. in begineer friendly way.

• go-auth-service: A Go service responsible for user authentication and authorization.
• go-blogs-service: A Go service responsible for managing blog posts.
• angular-blog-client: A Angular application that serves as the front-end client for the blog service.

Start the microservices (locally)

Build the Docker images

• Run Postgres db inside Docker (I don't have it natively)

sh +x scripts/postgres.sh

• Build Auth Service Microservice container

cd go-auth-service
make docker-build-push

• Build Blogs Service Microservice container

cd go-blogs-service
make docker-build-push

• Build Blogs Frontend Appservice container

cd angular-frontend
make docker-build-push

Images which I built: (just your reference)

Docker compose infrastructure:

Following script is present inside scripts/postgres.sh file. Which i built in terms of running the Postgresql in my machine. Later evolved to run the complete infrastructure.

PS: Felt lazy to recreate in a new Docker-compose yaml.

#!/bin/bash

# check the available version
docker --version
# check prune necessary setups
docker system prune
# create a docker volume
docker volume create --name=postgres-data

# run the postgres compose container
docker-compose -f postgres.docker-compose.yaml down
docker-compose -f postgres.docker-compose.yaml up --build

Sample docker runs

Note: The Docker images that i built, I kept in my private aartifacts registry. (Google-artifact-registry). In case you want to test it out locally use Dockerhub / local.docker repository.

Cloud SQL (Postgres) Instance

Try setup the proxy and KSA -> GSA Annotation into your postgres instance

Which will help your GKE services to communicate with postgres-cloudsql-instance

gcloud sql instances describe postgres-instance-demo

# Create Kubenetes secrets for cloud sql in GKE
kubectl create secret generic postgres-cloudsql-secret \
  --from-literal=username=root \
  --from-literal=password=some-good-password \
  --from-literal=authdb=auth \
  --from-literal=blogsdb=blogs \
  -n "fullstack-microservice-golang"

# Create a Kube SA account to connect to Cloud SQL via auth proxy
kubectl apply -f service-account.yaml

# -- to connect cloud sql from GKE
gcloud iam service-accounts add-iam-policy-binding \
--role="roles/iam.workloadIdentityUser" \
--member="serviceAccount:sounish-cloud-workstation.svc.id.goog[some-good-k8s-namespace/somegood-service-account]" \
some-good-service-account@sounish-cloud-workstation.iam.gserviceaccount.com

# Turn your service account key into a k8s Secret:
kubectl create secret generic cloudsql-connect-ksa-gsa \
--from-file=service_account.json=~/key.json

# Mount the secret as a volume under the spec: for your k8s object:
volumes:
- name: cloudsql-connect-ksa-gsa-vol
  secret:
    secretName: cloudsql-connect-ksa-gsa

Note: here, I am using auth-proxy-service, you can now use the workload-identity support for GKE.

Refer Documentation from Google: https://cloud.google.com/sql/docs/postgres/connect-kubernetes-engine#service-account-key

Kubernetes Deployments Infrastructure

All kubenetes manifests you can find in k8s-infrastructure/*.yaml directory.

• namespace.yaml - To have an isolated environment only for this microservices projects
• cloudsql-auth-proxy.yaml : CloudSQL (Postgres) is the backend database I am using. To access cloudsql from the GKE k8s-cluster, AUTHPROXY / Workloads identity must be created for security.
• auth-service.yaml : Contains the authentication microservice arch container
• blogs-service.yaml: Contains the blogs microservice arch container
• Blogs-app-frontend.yaml : Externally exposed frontend app service on Ngnix webserver.

• Google K8s Autopilot cluster:

• To create Autopilot Cluster:

#!/bin/bash

gcloud beta container --project "<YOUR_GCP_PROJECT_ID>" clusters create-auto "sounish-autopilot-cluster-01" --region "asia-south1" --release-channel "regular" --network "projects/<YOUR_GCP_PROJECT_ID>/global/networks/default" --subnetwork "projects/<YOUR_GCP_PROJECT_ID>/regions/asia-south1/subnetworks/default" --cluster-ipv4-cidr "/17" --binauthz-evaluation-mode=DISABLED

Bonus: If you are new to K8s then Try this Quick Hack

If you kind the manifests YAML hard to understand. You can learn the kubectl command line, to which you can generate your desireed boiler plates (QUICK HACK).

• Checkout scripts/kubernetes_service_deployments.sh file

#!/bin/bash

## based on the postgres.docker.compose.yaml file
# it has the kubernetes deployment manifests generation command line

# -- Auth-service deployment
kubectl create deployment auth-service --image=asia-south1-docker.pkg.dev/sounish-cloud-workstation/fullstack-microservices-golang-gke/go-auth-service:v241000.3127 --port=3000 --replicas=2 --output=yaml --show-managed-fields=true --namespace=fullstack-microservice-golang --dry-run=client

# -- Create a ClusterIP service for auth service 
kubectl create service clusterip auth-service --tcp=3000 --output=yaml --namespace=fullstack-microservice-golang --dry-run=client

# -- Blogs-service deployment
kubectl create deployment blog-service --image=asia-south1-docker.pkg.dev/sounish-cloud-workstation/fullstack-microservices-golang-gke/go-blogs-service:v241015.4818 --port=3001 --replicas=2 --output=yaml --show-managed-fields=true --namespace=fullstack-microservice-golang --dry-run=client

# -- Create a ClusterIP service for blog service 
kubectl create service clusterip blog-service --tcp=3001 --output=yaml --namespace=fullstack-microservice-golang --dry-run=client

# -- Blogs Frontend App deployment
kubectl create deployment blogs-app-frontend --image=asia-south1-docker.pkg.dev/sounish-cloud-workstation/fullstack-microservices-golang-gke/blogs-app-frontend:v241000.5219 --port=80 --replicas=2 --output=yaml --namespace=fullstack-microservice-golang --dry-run=client

# Blogs Frontend App Service
kubectl create service loadbalancer blogs-app-frontend --tcp=4200:80 --output=yaml --namespace=fullstack-microservice-golang --dry-run=client 

• K8s deployed workloads:

Sample Authservice - Workloads

• Auth service observbility Dashboard
• 
• Auth service Service Avaibility
• 
• Auth service Pods network
• 
• K8s Services and Ingress
• 

Sample blog content

With great power, comes great responsibility.
This blog talks about the scenarios.
Sounish being now GKE (google kubernetes engine).
Google cloud platform - Google
Thank you for your time.

Blogs App - GKE Microservices Demo

Login / Signup

Recommended Blogs - Home

Recommended Blogs - Home

Create new Blogpost

Kindly Appreciate If you Loved it

Made with ❤️ by Sounish Nath using Angular, Golang, Postgresql, Google Kubernetes Engine 🚀.