Wingie DevOps Case Study: GCP Kubernetes Project

This project involves setting up a scalable, observable, and modern microservice infrastructure using Kubernetes on Google Cloud Platform. The project includes creating a Kubernetes cluster using GKE (Google Kubernetes Engine), configuring custom node pools, deploying a sample application, implementing auto-scaling, monitoring, and setting up a service mesh.

Table of Contents

• Project Summary
• Architecture
• Infrastructure Features
• Application: Book Catalog API
• Monitoring and Alert System
• API Usage Examples
• Access Information
• Installation Steps

Project Summary

This case study demonstrates how to set up and configure a modern microservice infrastructure. The project includes the following main components:

• GKE Kubernetes Cluster: A Kubernetes cluster with customized node pools
• Book Catalog API: A sample PHP Laravel application
• Auto Scaling: Load-based scaling implemented with HPA and KEDA
• Monitoring System: Monitoring and alerting system implemented with Prometheus and Grafana
• Service Mesh: Secure and manageable service-to-service communication with Istio

Architecture

The project has a multi-layered architecture consisting of the following components visualized with interactive diagrams:

General Architecture Diagram

graph TD
    classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:white
    classDef infra fill:#4285f4,stroke:#fff,stroke-width:1px,color:white
    classDef monitoring fill:#fbbc04,stroke:#fff,stroke-width:1px,color:black
    classDef app fill:#34a853,stroke:#fff,stroke-width:1px,color:white
    classDef mesh fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white

    GKE["GKE Cluster<br>europe-west1"]:::k8s
    
    GKE --> MainPool["main-pool<br>n2d-standard-2<br>No scaling"]:::infra
    GKE --> AppPool["application-pool<br>n2d-standard-2<br>scaling: 1-3 nodes"]:::infra
    GKE --> MonitoringSystem["Monitoring System"]:::monitoring
    
    AppPool --> Frontend["Frontend<br>Nginx"]:::app
    AppPool --> Backend["Backend<br>PHP-FPM Laravel"]:::app
    AppPool --> DB["Database<br>PostgreSQL"]:::app
    
    MonitoringSystem --> Prometheus["Prometheus<br>Metrics Collection"]:::monitoring
    MonitoringSystem --> Grafana["Grafana<br>Dashboard & Alerts"]:::monitoring
    
    Prometheus -.-> DB
    Prometheus -.-> Backend
    Prometheus -.-> Frontend
    
    IstioDeploy["Istio Service Mesh"]:::mesh
    GKE --> IstioDeploy
    
    IstioDeploy --> Istiod["istiod<br>control plane"]:::mesh
    IstioDeploy --> IngressGateway["istio-ingressgateway"]:::mesh
    IstioDeploy --> EgressGateway["istio-egressgateway"]:::mesh
    
    IngressGateway --> Frontend
    
    KEDA["KEDA<br>AutoScaler"]:::infra
    GKE --> KEDA
    KEDA -.-> Backend
    
    HPA["Horizontal Pod Autoscaler<br>CPU > 25%"]:::infra
    GKE --> HPA
    HPA -.-> Backend
    HPA -.-> Frontend

Loading

Node Pool Structure and Deployment Relationship

flowchart TD
    classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:white
    classDef infra fill:#4285f4,stroke:#fff,stroke-width:1px,color:white
    classDef app fill:#34a853,stroke:#fff,stroke-width:1px,color:white
    classDef selector fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white

    GKE["GKE Cluster"]:::k8s
    
    GKE --> MainPool["main-pool<br>n2d-standard-2<br>Auto-scaling: Disabled"]:::infra
    GKE --> AppPool["application-pool<br>n2d-standard-2<br>Auto-scaling: 1-3 nodes"]:::infra
    
    AppPool --> NodeSelector["nodeSelector:<br>pool=application-pool"]:::selector
    
    NodeSelector --> FrontendPod["Frontend Pod<br>Nginx"]:::app
    NodeSelector --> BackendPod["Backend Pod<br>PHP-FPM Laravel"]:::app
    NodeSelector --> DBPod["Database Pod<br>PostgreSQL"]:::app

Loading

Monitoring and Alerting Structure

graph TD
    classDef monitoring fill:#fbbc04,stroke:#fff,stroke-width:1px,color:black
    classDef alert fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white
    classDef metrics fill:#4285f4,stroke:#fff,stroke-width:1px,color:white

    Prometheus["Prometheus<br>Time Series DB"]:::monitoring
    
    Prometheus --> PrometheusOperator["Prometheus Operator"]:::monitoring
    PrometheusOperator --> ServiceMonitor["ServiceMonitor<br>Metrics Scraping"]:::monitoring
    PrometheusOperator --> PrometheusRule["PrometheusRule<br>Alert Rules"]:::monitoring
    
    ServiceMonitor --> AppMetrics["Application Metrics<br>CPU, Memory, Requests"]:::metrics
    ServiceMonitor --> K8sMetrics["Kubernetes Metrics<br>Node, Pod Status"]:::metrics
    
    PrometheusRule --> PodRestartAlert["Pod Restart Alert<br>15 minutes restart window"]:::alert
    PrometheusRule --> CPUAlert["CPU Usage Alert<br>>25% CPU"]:::alert
    
    Prometheus --> Grafana["Grafana<br>Dashboard & Alerts"]:::monitoring
    Grafana --> Dashboard["Kubernetes Dashboard"]:::monitoring
    Grafana --> Alerts["Alert Management"]:::monitoring

Loading

Auto-Scaling Structure: HPA and KEDA Comparison

graph TD
    classDef k8s fill:#326ce5,stroke:#fff,stroke-width:1px,color:white
    classDef autoscaler fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white
    classDef app fill:#34a853,stroke:#fff,stroke-width:1px,color:white
    classDef metrics fill:#fbbc04,stroke:#fff,stroke-width:1px,color:black

    K8s["Kubernetes Cluster"]:::k8s
    
    K8s --> HPA["Horizontal Pod Autoscaler"]:::autoscaler
    K8s --> KEDA["Kubernetes Event-Driven Autoscaler"]:::autoscaler
    
    HPA --> HPAMetrics["CPU Usage > 25%<br>1-3 pods"]:::metrics
    HPAMetrics --> HPADeployment["Book Catalog Web/App<br>Deployments"]:::app
    
    KEDA --> KEDAMetrics["CPU Usage > 25%<br>1-3 pods"]:::metrics
    KEDA --> ScaledObject["ScaledObject<br>book-catalog-scaled-object"]:::autoscaler
    ScaledObject --> BookCatalog["Book Catalog<br>Deployment"]:::app
    
    %% Note: While KEDA is shown with CPU-based scaling in this example,
    %% it can also scale based on other data sources (Prometheus, Kafka, RabbitMQ, etc.)

Loading

Istio Service Mesh Architecture

graph TD
    classDef mesh fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white
    classDef app fill:#34a853,stroke:#fff,stroke-width:1px,color:white
    classDef traffic fill:#4285f4,stroke:#fff,stroke-width:1px,color:white
    classDef security fill:#fbbc04,stroke:#fff,stroke-width:1px,color:black

    Istio["Istio Service Mesh"]:::mesh
    
    Istio --> Istiod["istiod<br>Control Plane"]:::mesh
    Istio --> IngressGW["istio-ingressgateway"]:::mesh
    Istio --> EgressGW["istio-egressgateway"]:::mesh
    
    IngressGW --> IstioGateway["Istio Gateway<br>Ingress Rules"]:::traffic
    IstioGateway --> VirtualService["VirtualService<br>Traffic Rules"]:::traffic
    
    VirtualService --> Frontend["Frontend Service"]:::app
    Frontend --> Backend["Backend Service"]:::app
    Backend --> DB["Database Service"]:::app
    
    Istiod --> mTLS["Mutual TLS<br>Service Security"]:::security
    Istiod --> Telemetry["Telemetry<br>Tracing, Metrics"]:::traffic

Loading

Application Architecture: Book Catalog API

graph LR
    classDef infra fill:#4285f4,stroke:#fff,stroke-width:1px,color:white
    classDef app fill:#34a853,stroke:#fff,stroke-width:1px,color:white
    classDef db fill:#326ce5,stroke:#fff,stroke-width:1px,color:white
    classDef api fill:#ff6d00,stroke:#fff,stroke-width:1px,color:white

    User(("User")) --> Ingress["LoadBalancer<br>External IP"]:::infra
    
    Ingress --> FrontendSvc["Frontend Service"]:::infra
    FrontendSvc --> Frontend["Nginx<br>Web Server"]:::app
    
    Frontend --> BackendSvc["Backend Service"]:::infra
    BackendSvc --> Backend["PHP-FPM<br>Laravel API"]:::app
    
    Backend --> DBSvc["Database Service"]:::infra
    DBSvc --> DB["PostgreSQL<br>Book Catalog DB"]:::db
    
    Backend --> API["RESTful API Endpoints"]:::api
    API --> ListBooks["GET /api/books"]:::api
    API --> GetBook["GET /api/books/{id}"]:::api
    API --> CreateBook["POST /api/books"]:::api
    API --> UpdateBook["PUT /api/books/{id}"]:::api
    API --> DeleteBook["DELETE /api/books/{id}"]:::api

Loading

Infrastructure Features

GKE Cluster

• Region: europe-west1
• Logging and Monitoring: Disabled
• Kubernetes Version: 1.27.3-gke.100

Node Pools

1. main-pool

   • Machine Type: n2d-standard-2
   • Auto Scaling: Disabled (fixed 1 node)
   • Labels: pool=main-pool

2. application-pool

   • Machine Type: n2d-standard-2
   • Auto Scaling: Enabled (min:1, max:3 nodes)
   • Labels: pool=application-pool

Application Deployment

• Book Catalog Application: Microservice architecture with 3 separate deployments
  • Frontend (Nginx)
  • Backend (PHP-FPM)
  • Database (PostgreSQL)
• NodeSelector: All applications run on application-pool

Auto Scaling

• HPA: Scales between 1-3 pods when CPU usage exceeds 25%
• KEDA: Alternative scaling from 1-3 pods when CPU usage exceeds 25%

Monitoring System

• Prometheus: Collecting Kubernetes metrics
• Grafana: Visualization and alert configuration
• Pod Restart Alert: Alerts when any pod restarts within 15 minutes

Service Mesh

• Istio: Mesh structure for inter-service communication
  • istiod (control plane)
  • istio-ingressgateway
  • istio-egressgateway

Application: Book Catalog API

Book Catalog API is a simple CRUD application for managing a book catalog. It was developed using Laravel and PostgreSQL.

Application Features

• API Based: Book management via RESTful API
• Database: PostgreSQL
• Architecture: Laravel MVC
• Language: PHP 8.2

Book Model

Each book has the following attributes:

• id: Unique identifier
• name: Title of the book
• author: Author of the book
• pages: Number of pages
• created_at: Date added
• updated_at: Last update date

API Usage Examples

List all books

curl http://35.241.242.66/api/books

Add a new book

curl -X POST http://35.241.242.66/api/books \
  -H "Content-Type: application/json" \
  -d '{"name":"1984","author":"George Orwell","pages":328}'

View a specific book

curl http://35.241.242.66/api/books/1

Update a book

curl -X PUT http://35.241.242.66/api/books/1 \
  -H "Content-Type: application/json" \
  -d '{"name":"1984","author":"George Orwell","pages":352}'

Delete a book

curl -X DELETE http://35.241.242.66/api/books/1

Monitoring and Alert System

Prometheus

• Features: Time series database, query language, alert system
• Data Collection: Kubernetes metrics, pod status, CPU and memory usage
• Scrape Interval: 15 seconds

Grafana

• Features: Visualization, dashboards, alerts
• Dashboards: Kubernetes overview, pod status, resource usage
• Alerts:
  • Pod restart alert (any restart within 15 minutes)
  • Pod high CPU usage
  • Pod high memory usage

Access Information

Book Catalog API

• External Access URL: http://35.241.242.66/api/books
• Service Type: LoadBalancer
• Port: 80

Prometheus

• URL: http://34.77.133.67
• Service Type: LoadBalancer

Grafana

• URL: http://35.205.72.133
• Service Type: LoadBalancer
• Username/Password: admin/admin

Installation Steps

Steps required for project installation:

1. Setting up GCP Infrastructure with Terraform

cd infrastructure/terraform
terraform init
terraform apply

2. Applying Kubernetes Manifests

kubectl apply -f infrastructure/k8s/book-catalog/deployment.yaml
kubectl apply -f infrastructure/k8s/book-catalog/service.yaml
kubectl apply -f infrastructure/k8s/book-catalog/hpa.yaml
kubectl apply -f infrastructure/k8s/book-catalog/nginx-configmap.yaml
kubectl apply -f infrastructure/k8s/book-catalog/migration-job.yaml

3. Installing Prometheus and Grafana

helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
helm repo update
helm install prometheus prometheus-community/kube-prometheus-stack -n monitoring --create-namespace -f infrastructure/monitoring/prometheus-values.yaml
kubectl apply -f infrastructure/monitoring/prometheus-services.yaml
kubectl apply -f infrastructure/monitoring/pod-restart-rule.yaml
kubectl apply -f infrastructure/monitoring/book-catalog-servicemonitor.yaml

4. Installing KEDA

helm repo add kedacore https://kedacore.github.io/charts
helm repo update
helm install keda kedacore/keda --namespace keda --create-namespace
kubectl apply -f infrastructure/k8s/book-catalog/keda-scaledobject.yaml

5. Installing Istio

curl -L https://istio.io/downloadIstio | ISTIO_VERSION=1.18.0 sh -
cd istio-1.18.0
./bin/istioctl install --set profile=demo -y
kubectl label namespace default istio-injection=enabled
kubectl apply -f infrastructure/k8s/book-catalog/istio-gateway.yaml

---

This project was prepared to demonstrate DevOps and Cloud Native application development practices. It shows step by step how to set up and configure a modern microservice infrastructure using Kubernetes on GKE.