@webalexx/projen-multi-cloud-project

A TypeScript Projen library for creating multi-cloud Pulumi projects with AWS EKS, Minikube, and FastAPI ML Services.

Features

• 🚀 FastAPI ML Services: Pre-configured FastAPI templates with LLM integration (AWS Bedrock)
• ☸️ Kubernetes Deployment: Support for Minikube and AWS EKS
• 🐳 Docker Multi-Stage Builds: Optimized container images for production
• 📊 Helm Charts: Complete Helm chart templates with HPA and Secrets management
• 🔐 AWS Secrets Manager: Native integration for secure credential management
• 🔄 GitHub Actions CI/CD: Complete workflows for testing, building, and deployment
• 🧪 Unit & Integration Tests: Pytest for Python services + Jest for TypeScript

Quick Start

Installation

npm install @webalexx/projen-multi-cloud-project --save-dev

Create .projenrc.ts

import { MultiCloudPulumiProject, MLProject } from '@webalexx/projen-multi-cloud-project';

const project = new MultiCloudPulumiProject({
  name: '@myorg/my-ml-project',
  repositoryUrl: 'https://github.com/myorg/my-ml-project.git',
  author: 'Your Name',
  authorAddress: '[email protected]',
  defaultReleaseBranch: 'main',
});

// Create an ML service
new MLProject(project, 'data-analyzer', {
  projectName: 'data-analyzer',
  awsRegion: 'us-east-1',
  enableMinikube: true,
  enableEks: true,
  enableHelmChart: true,
});

project.synth();

Generate Project

npx projen

API Reference

MultiCloudPulumiProject

High-level construct that creates a Projen TypeScript project with multi-cloud Pulumi support.

| Class / Method | Description | Parameters | | :--- | :--- | :--- | | MultiCloudPulumiProject | Main project class combining all cloud components | options: MultiCloudPulumiProjectOptions | | MultiCloudPulumiProjectOptions.author | Author name | string | | MultiCloudPulumiProjectOptions.authorAddress | Author email address | string | | MultiCloudPulumiProjectOptions.repositoryUrl | Git repository URL | string |

MLProject

High-level construct that generates a complete ML service with FastAPI, Docker, Helm, and Pulumi infrastructure.

| Class / Method | Description | Parameters | | :--- | :--- | :--- | | MLProject | Creates a complete ML service component | scope: TypeScriptProject, id: string, options: MLProjectOptions | | MLProjectOptions.projectName | Name of the ML service (lowercase alphanumeric and hyphens, e.g. data-analyzer) | string (required) | | MLProjectOptions.projectDescription | Human-readable service description | string (default: "ML Service: <projectName>") | | MLProjectOptions.awsRegion | AWS region for deployment | string (default: "us-east-1") | | MLProjectOptions.awsAccountId | AWS account ID for ECR image path | string (optional) | | MLProjectOptions.llmProvider | LLM provider to use | string (default: "bedrock") | | MLProjectOptions.imageRepository | Docker image repository name | string (default: "<projectName>-service") | | MLProjectOptions.enableHelmChart | Generate Helm chart for Kubernetes deployment | boolean (default: true) | | MLProjectOptions.enableMinikube | Generate GitHub Actions workflow for Minikube | boolean (default: true) | | MLProjectOptions.enableEks | Generate GitHub Actions workflow for AWS EKS | boolean (default: true) |

Generated Files

When you instantiate MLProject, the following files are generated in your project:

| File | Description | | :--- | :--- | | services/<name>/main.py | FastAPI application with AWS Bedrock LLM integration | | services/<name>/requirements.txt | Python dependencies | | services/<name>/test_main.py | Pytest unit and integration tests | | services/<name>/.env.example | Environment variable template (12-factor) | | services/<name>/Dockerfile | Multi-stage production Docker image | | services/<name>/.dockerignore | Docker build exclusions | | infrastructure/pulumi/<name>-eks.ts | Pulumi code for AWS EKS cluster provisioning | | infrastructure/pulumi/<name>-minikube.ts | Pulumi code for Minikube local deployment | | helm-charts/<name>/Chart.yaml | Helm chart metadata | | helm-charts/<name>/values.yaml | Helm chart default values | | helm-charts/<name>/templates/deployment.yaml | Kubernetes Deployment manifest | | helm-charts/<name>/templates/service.yaml | Kubernetes Service manifest | | .github/workflows/deploy-minikube-<name>.yml | CI/CD workflow for Minikube (if enableMinikube: true) | | .github/workflows/deploy-eks-<name>.yml | CI/CD workflow for AWS EKS (if enableEks: true) |

API Endpoints

Each generated FastAPI service provides:

| Endpoint | Method | Description | | :--- | :--- | :--- | | /health | GET | Health check endpoint | | /info | GET | Service information (name, region, provider) | | /analyze | POST | Analyze and rate content using LLM |

Example

# Health check
curl http://localhost:8000/health

# Analyze content
curl -X POST http://localhost:8000/analyze \
  -H "Content-Type: application/json" \
  -d '{"text": "Your content here..."}'

Local Development

Docker

cd services/data-analyzer
docker build -t data-analyzer:latest .
# Use AWS IAM roles or OIDC for production; credentials are fetched at runtime
# from AWS Secrets Manager – do NOT pass secret keys as environment variables.
docker run -p 8000:8000 \
  -e AWS_REGION=us-east-1 \
  data-analyzer:latest

Minikube

minikube start
eval $(minikube docker-env)
docker build -t data-analyzer:latest services/data-analyzer/
cd infrastructure/pulumi
pulumi stack init dev-minikube
pulumi up

AWS EKS Deployment

# Using Helm
helm upgrade --install data-analyzer ./helm-charts/data-analyzer \
  --namespace data-analyzer-prod \
  --create-namespace

# Using Pulumi
cd infrastructure/pulumi
pulumi stack init prod-eks
pulumi config set target eks
pulumi up

Running Tests

# TypeScript library tests
npm test

# Python service tests
pytest services/data-analyzer/test_main.py -v --cov

12-Factor Compliance

All generated services follow 12-factor app principles:

• Configuration via environment variables (not hardcoded)
• Secrets fetched from AWS Secrets Manager at runtime
• Stateless processes with no local state

License

Apache 2.0