| name | azure-architect |
| description | Azure architecture design: App Service, AKS, Azure Functions, Cosmos DB, Bicep IaC templates, cost optimisation, Entra ID integration, and migration patterns |
Azure Architect Skill
When to activate
- Designing a new Azure architecture from scratch
- Choosing between Azure service options (App Service vs AKS vs Functions, Cosmos DB vs Azure SQL, etc.)
- Generating Bicep or ARM templates for a pattern
- Optimising Azure costs on an existing deployment
- Planning a migration to Azure from on-prem or another cloud
- Setting up Azure DevOps or GitHub Actions for Azure deployments
When NOT to use
- AWS-specific architecture — use the aws-architect skill
- GCP-specific architecture — use the gcp-architect skill
- Cloud security posture review — use the cloud-security skill
Instructions
Architecture pattern selection
Select the right Azure architecture pattern for [application].
Application type: [web app / mobile backend / SaaS / microservices / data pipeline]
Scale: [users/day, requests/second]
Team Azure experience: [beginner / intermediate / advanced]
Budget: $[X]/month target
Compliance: [GDPR / HIPAA / ISO 27001 / PCI / none]
Microsoft stack: [already using M365 / Teams / Azure AD? Yes / No]
Azure pattern guide:
APP SERVICE (recommended for: web apps, APIs, low ops overhead):
Stack: Front Door + App Service + Azure SQL + Redis Cache + Key Vault
Cost: ~$50-300/month depending on tier
Pros: familiar PaaS, built-in deployment slots (blue-green), Entra ID integration
Cons: less flexible than containers, platform-imposed limits
Best for: .NET, Node.js, Python web apps; teams migrating from IIS or other PaaS
AZURE FUNCTIONS (recommended for: event-driven, serverless):
Stack: Front Door + Functions + Cosmos DB + Service Bus + Storage
Cost: ~$0-50/month for small workloads (Consumption plan)
Pros: pay-per-execution, auto-scale, 200+ triggers/bindings
Cons: cold starts on Consumption plan; avoid for > 10 min execution; stateful needs Durable Functions
Best for: event-driven microservices, background jobs, webhooks, scheduled tasks
AKS (recommended for: microservices, existing K8s teams):
Stack: Front Door + AKS + Azure SQL / Cosmos DB + Service Bus + ACR
Cost: ~$200-1000/month minimum (running node pools 24/7)
Pros: full Kubernetes control, multi-cloud portability, KEDA for scaling
Cons: operational complexity, higher baseline cost
Best for: teams with K8s experience, complex microservices, needing fine-grained control
DATA / ANALYTICS:
Stack: Event Hubs + Stream Analytics / Databricks + Azure Data Lake + Synapse Analytics
Best for: IoT, telemetry, data warehousing, BI
Recommend the pattern for my application with cost estimate and Bicep starter template.
