Scalable Object Identification App

This repository demonstrates a scalable object identification application, emphasizing scalability and monitoring. The application is containerized using Docker and deployed on a local Kubernetes cluster via Minikube. Monitoring is integrated using Prometheus, and a load testing script is provided to evaluate performance under varying workloads.

Features

• Scalability: Designed to efficiently handle increasing workloads within a Kubernetes environment.

• Containerization: Utilizes Docker for consistent and portable deployment across different environments.

• Kubernetes Deployment: Deployed on Minikube, facilitating local development and testing of Kubernetes applications.

• Monitoring: Integrated with Prometheus to collect and visualize performance metrics, aiding in proactive system management.

• Load Testing: Includes scripts to simulate various load scenarios, enabling performance evaluation and optimization.

Prerequisites

Before setting up the environment, ensure you have the following installed:

• Python 3

• Docker

• Minikube

• kubectl

• Helm

Setup Environment

1. Load the Model:

   python3 scripts/load_model.py

2. Configure Docker to Use Minikube's Environment:

   eval $(minikube docker-env)

3. Build Docker Images:

   • Flask Application:

     docker buildx build -t flask-app:latest -f ./docker/producer/Dockerfile.flaskapp .

   • Worker Application:

     docker buildx build -t worker-app:latest -f ./docker/consumer/Dockerfile.worker .

   • Custom Autoscaler:

     docker buildx build -t autoscaler:latest . -f Dockerfile.CPA

4. Start Minikube:

   minikube start --driver=docker

   Note: You can specify the number of CPU cores and memory for Minikube using the --cpus and --memory flags. For example:

   minikube start --driver=docker --cpus 4 --memory 8192

Setting Up Prometheus

Prometheus is used for monitoring the application's performance metrics.

1. Install Prometheus Using Helm:

   • Add Helm Repository:

     helm repo add prometheus-community https://prometheus-community.github.io/helm-charts
     helm repo update

   • Create Monitoring Namespace:

     kubectl apply -f ./deployments/monitoring-namespace.yaml

   • Install Prometheus:

     helm install prometheus prometheus-community/kube-prometheus-stack --namespace monitoring

     This command installs Prometheus and related components into the monitoring namespace.

2. Apply Prometheus Custom Resource Definitions (CRDs):

   chmod +x scripts/apply-prometheus-crds.sh
   ./scripts/apply-prometheus-crds.sh

3. Upgrade Prometheus with Custom Configuration:

   helm upgrade prometheus prometheus-community/kube-prometheus-stack -f deployments/prometheus.yaml -n monitoring

Deploying the Flask Application

Deploy the Flask application using the provided Kubernetes manifests:

kubectl apply -f deployments/app.yaml

Deploying the Custom Autoscaler

1. Ensure the Default Horizontal Pod Autoscaler (HPA) Is Not Active:

   kubectl delete hpa flask-app-hpa

2. Deploy the Custom Autoscaler:

   kubectl apply -f ./deployments/autoscaler.yaml

Accessing the Deployment

To access the deployed application:

1. Start Minikube Tunnel:

   minikube tunnel

2. Verify Deployment:

   • Open a web browser and navigate to http://localhost:5000 to access the application.

Monitoring and Visualization

• Access Minikube Dashboard:

  minikube dashboard

This command opens the Kubernetes dashboard, providing an overview of the cluster's resources and deployments.

• Access Prometheus UI:

  kubectl port-forward service/prometheus-kube-prometheus-prometheus 9090:9090 -n monitoring

Navigate to http://localhost:9090 in your web browser to access the Prometheus interface.

• Access Grafana UI:

  kubectl port-forward -n monitoring svc/prometheus-grafana 3000:80

Open http://localhost:3000 and log in with the default credentials:

• Username: admin

• Password: prom