Skip to content
/ backend-deploy Public

PaletAI Backend Azure deployment infrastructure with cost optimization (B1 SKU - 3/month). Complete migration toolkit for deploying to new Azure subscription/tenant.

0 stars 0 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

HuddleUp-AI/backend-deploy

BranchesTags

Repository files navigation

PaletAI Backend - Azure Deployment Repository

Complete Azure infrastructure deployment for PaletAI Backend with cost optimization and migration tools.

🎯 Purpose

This repository contains everything needed to deploy PaletAI Backend to a new Azure subscription/tenant with minimal cost configuration (B1 App Service Plan - $13/month).

Use Case: Migrating from old Azure subscription to new subscription/tenant while reducing costs by 70-85%.

📁 Repository Contents

File Purpose Audience
PM-INTRODUCTION.md Non-technical overview & business case 📋 Product Managers
MIGRATION-GUIDE.md Complete step-by-step migration guide 🔧 DevOps Engineers
MIGRATION-CHECKLIST.md Day-of-migration execution checklist 🔧 DevOps Engineers
Deploy-PaletAI.ps1 Automated deployment script 🔧 DevOps Engineers
main.bicep Azure infrastructure template (IaC) 🔧 DevOps Engineers
QUICK-REFERENCE.md Common commands reference card 🔧 DevOps Engineers
env-template.txt Environment variables reference 🔧 DevOps Engineers
parameters.json Default deployment parameters 🔧 DevOps Engineers
README.md Repository overview 👥 Everyone

🚀 Quick Start

Prerequisites

  • Azure PowerShell Module (Install-Module -Name Az) or Azure CLI
  • Contributor/Owner access to target Azure subscription
  • MongoDB connection string
  • AI provider API key (OpenAI, Azure OpenAI, or Anthropic)

Deploy in 3 Commands

# 1. Clone this repository
git clone https://github.com/HuddleUp-AI/backend-deploy.git
cd backend-deploy

# 2. Run deployment script
.\Deploy-PaletAI.ps1 `
  -TenantId "YOUR_TENANT_ID" `
  -SubscriptionId "YOUR_SUBSCRIPTION_ID" `
  -ResourceGroupName "rg-paletai-prod" `
  -MongoDbConnectionString "mongodb://..." `
  -OpenAiApiKey "sk-..." `
  -AiProvider "openai"

# 3. Deploy application code (see MIGRATION-GUIDE.md)

Total Time: ~8 minutes (5 min infrastructure + 3 min application)

💰 Cost Savings

Environment Monthly Cost Savings
Old Subscription $70-100/month -
New Subscription (B1) $15-16/month $55-85/month

Annual Savings: $660-$1,020/year (70-85% reduction)

📖 Documentation

For Product Managers

  • PM-INTRODUCTION.md - 📋 Non-technical overview - Business value, timeline, risks, stakeholder communication

Getting Started (Technical)

  1. README.md - Quick overview and deployment scenarios
  2. MIGRATION-GUIDE.md - Comprehensive migration guide (START HERE for engineers)
  3. MIGRATION-CHECKLIST.md - Day-of-migration execution checklist

Reference

🏗️ What Gets Deployed

Azure Resources

  • App Service Plan (B1) - Linux, Python 3.12

    • 1 vCPU, 1.75GB RAM
    • AlwaysOn enabled
    • Custom domain support

  • App Service (Web App) - PaletAI Backend API

    • Python 3.12 runtime
    • Gunicorn + Uvicorn (ASGI)
    • HTTPS only, TLS 1.2+
    • Configured for FastAPI

  • Storage Account (LRS) - Game image storage

    • Blob container: game-images
    • Public blob access enabled
    • CORS configured for web access

  • Application Insights - Monitoring and diagnostics

    • Live metrics
    • Performance tracking
    • Error logging

Estimated Monthly Costs

Resource SKU Monthly Cost
App Service Plan B1 $13.00
Storage Account Standard LRS $1-2.00
Application Insights Pay-as-you-go $0-1.00 (5GB free)
Total ~$15-16

⚙️ Configuration

Supported Azure Regions

  • westus3 (default - newest, cost-effective)
  • eastus (established, highly available)
  • westus2, centralus, eastus2, etc.

App Service Plan SKUs

SKU Monthly Cost Use Case
B1 $13 ✅ Recommended for most workloads
B2 $26 Higher traffic, more CPU/RAM
B3 $52 High performance requirements
S1 $70 Auto-scaling, staging slots
P1V2 $140 Premium performance

AI Provider Options

  • OpenAI (GPT-4o, GPT-4 Turbo, GPT-3.5)
  • Azure OpenAI (GPT-4o, GPT-4)
  • Anthropic (Claude 3.5 Sonnet, Claude 3 Opus)

Choose provider via -AiProvider parameter.

🔧 Common Operations

Deploy to Different Region

.\Deploy-PaletAI.ps1 `
  -TenantId "..." `
  -SubscriptionId "..." `
  -ResourceGroupName "rg-paletai-eastus" `
  -Location "eastus" `
  -MongoDbConnectionString "..." `
  -OpenAiApiKey "..."

Deploy Development Environment

.\Deploy-PaletAI.ps1 `
  -TenantId "..." `
  -SubscriptionId "..." `
  -ResourceGroupName "rg-paletai-dev" `
  -Environment "dev" `
  -AppServicePlanSku "B1" `
  -MongoDbConnectionString "..." `
  -AnthropicApiKey "..." `
  -AiProvider "anthropic"

Scale Up/Down

# Scale to B2 (2 vCPU, 3.5GB RAM)
az appservice plan update \
  --name PLAN_NAME \
  --resource-group rg-paletai-prod \
  --sku B2

# Scale back to B1
az appservice plan update \
  --name PLAN_NAME \
  --resource-group rg-paletai-prod \
  --sku B1

📋 Migration Process Overview

  1. Pre-Migration (1-2 days before)

    • Gather credentials and connection strings
    • Backup MongoDB database
    • Test connectivity from new region
    • Lower DNS TTL (if using custom domain)

  2. Deploy Infrastructure (~5 minutes)

    • Run Deploy-PaletAI.ps1 script
    • Verify resources created in Azure Portal

  3. Deploy Application (~3 minutes)

    • GitHub Actions (recommended) or ZIP deploy
    • Verify health endpoint

  4. Database Migration (varies)

    • Same MongoDB: No migration needed
    • New MongoDB: Backup/restore
    • Zero downtime: Use continuous sync

  5. DNS Cutover (if applicable)

    • Update CNAME to new app service
    • Wait for propagation (5-60 minutes)

  6. Verification (~30 minutes)

    • Test all endpoints
    • Monitor Application Insights
    • Verify user functionality

  7. Decommission Old Environment (after 24-48 hours)

    • Stop old app service
    • Delete after 7-30 days

See MIGRATION-GUIDE.md for detailed instructions.

🔐 Security

  • ✅ All secrets passed as secure parameters
  • ✅ HTTPS enforced for all connections
  • ✅ TLS 1.2 minimum version
  • ✅ No secrets in version control
  • ✅ Application Insights for monitoring
  • ✅ Managed identities support (future)

Warning: Never commit filled parameters.json or .env files with real secrets to git!

🆘 Troubleshooting

Deployment Fails

# Check deployment logs
az deployment group show \
  --name DEPLOYMENT_NAME \
  --resource-group rg-paletai-prod

# Retry with verbose logging
.\Deploy-PaletAI.ps1 ... -Verbose

Health Endpoint Returns 500

# Stream application logs
az webapp log tail \
  --name APP_NAME \
  --resource-group rg-paletai-prod

# Check app settings (especially DB_PATH)
az webapp config appsettings list \
  --name APP_NAME \
  --resource-group rg-paletai-prod

Database Connection Fails

  • Verify MongoDB connection string is correct
  • Check MongoDB allows connections from Azure IPs
  • Test connection with mongosh from local machine
  • Verify connection string in app settings

Full troubleshooting guide: MIGRATION-GUIDE.md#troubleshooting

🔄 Updates and Maintenance

Update Application Code

Recommended: Use GitHub Actions

  1. Get publish profile: az webapp deployment list-publishing-profiles ...
  2. Add to GitHub secrets as AZURE_WEBAPP_PUBLISH_PROFILE
  3. Push to main branch → auto-deploys

Alternative: ZIP deploy

az webapp deployment source config-zip \
  --resource-group rg-paletai-prod \
  --name APP_NAME \
  --src deploy.zip

Update Configuration

# Update environment variables
az webapp config appsettings set \
  --name APP_NAME \
  --resource-group rg-paletai-prod \
  --settings DAILY_PROMPT_LIMIT=50

# Restart app service
az webapp restart \
  --name APP_NAME \
  --resource-group rg-paletai-prod

📊 Monitoring

Application Insights

Access via: Azure Portal → Application Insights → YOUR_APP_NAME

Key metrics:

  • Live Metrics (real-time requests)
  • Failures (exceptions and errors)
  • Performance (response times)
  • Availability (uptime monitoring)

Cost Monitoring

# Create budget alert
az consumption budget create \
  --budget-name paletai-monthly \
  --amount 50 \
  --time-grain monthly \
  --resource-group rg-paletai-prod

🤝 Contributing

This repository is maintained by the HuddleUp-AI team. For issues or improvements:

  1. Create an issue in this repository
  2. Submit a pull request with proposed changes
  3. Contact the DevOps team

📞 Support

📜 License

Internal use for HuddleUp-AI projects.

Ready to deploy? Start with MIGRATION-GUIDE.md!

About

PaletAI Backend Azure deployment infrastructure with cost optimization (B1 SKU - 3/month). Complete migration toolkit for deploying to new Azure subscription/tenant.

Topics

devops azure infrastructure-as-code azure-deployment app-service cost-optimization fastapi bicep azure-infrastructure paletai

Resources

Readme
Activity
Custom properties

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages