azure-backend-architecture

name	azure-backend-architecture
description	Use this skill whenever backend modernization, strangler-backend, event-driven refactoring, cloud-native transition, Azure services, Azure Architecture Center, Well-Architected Framework, Azure MCP, AKS, Container Apps, App Service, Functions, Service Bus, Event Grid, API Management, or Azure backend best practices are mentioned. It gates backend target architecture before implementation.

Azure Backend Architecture

Backend modernization in this workflow targets Azure-based services. This skill turns the as-is backend and integration seams into a to-be Azure architecture that is grounded in Azure Architecture Center guidance, Well-Architected Framework pillars, and available Azure tooling/MCP context.

When to use

• Before strangler-backend code mode.
• Whenever the backend target architecture, Azure service choice, event-driven migration, DB2DB/TCP/API replacement, cloud-native transition, Azure MCP, AKS, Container Apps, App Service, Functions, Service Bus, Event Grid, API Management, observability, or Azure best practices are discussed.

Required inputs

• docs/renovation/02-architecture.md — as-is architecture, integration mechanics, hotspots, and seam inventory.
• docs/renovation/03-requirements.md — grilled requirements and target constraints.
• Azure organization constraints if available: landing zone, networking, identity, security, subscription/resource-group standards, cost constraints.

Azure guidance to consult

Use the reference summary in AZURE_REFERENCE.md and, when current details matter, fetch the live Microsoft Learn pages:

• Azure Application Architecture Fundamentals
• Azure Well-Architected Framework
• Azure Architecture Center best practices, especially API design
• Azure compute decision tree
• Azure cloud design patterns, especially messaging/event-driven patterns

Existing Azure skills to consider

The open skills ecosystem search found Microsoft Azure skills that may be useful. Reference or install them when the task needs that depth:

• microsoft/azure-skills@azure-enterprise-infra-planner
• microsoft/azure-skills@azure-kubernetes
• microsoft/azure-skills@azure-cost
• microsoft/azure-skills@azure-upgrade
• microsoft/azure-skills@azure-quotas

Do not duplicate those skills. Use this skill to route and frame the Azure backend architecture decisions for this renovation workflow.

Azure MCP / tooling rule

If Azure MCP tools are available in the current agent environment, use them to inspect real Azure resources, subscriptions, deployed services, networking, identity, and observability posture. If Azure MCP is not available, use Azure CLI (az) if logged in and safe, plus Microsoft Learn guidance. If neither is available, document the missing environment access and proceed in design mode only.

Run the helper if useful:

.agents/skills/azure-backend-architecture/scripts/check-azure-tooling.sh

Method

1. Map each backend seam to Azure candidates. For each seam from architecture-recovery, identify whether the target is API, event, worker, data store, cache, integration, or observability.
2. Choose compute deliberately. Use Azure Architecture Center compute guidance: App Service for managed web/API hosting, Azure Functions for event-driven functions, Container Apps for managed serverless containers, AKS when Kubernetes control-plane access and orchestration are required. Avoid defaulting to AKS without a need.
3. Prefer loose coupling and events. Replace DB2DB/TCP tight coupling with APIs, Service Bus/Event Grid/Event Hubs, or event-sourced read models where the use case needs async processing or realtime status.
4. Apply Well-Architected pillars. Evaluate reliability, security, cost optimization, operational excellence, and performance efficiency. Record tradeoffs.
5. Design API boundaries. APIs must be resource-oriented, loosely coupled, not database-table mirrors, and avoid chatty patterns.
6. Define observability. Every Azure slice needs logs, metrics, traces, alerts, dashboards, and SLO/SLA indicators. Prefer Azure Monitor/Application Insights/Log Analytics where applicable.
7. Create the Azure migration slice plan. Feed service decisions into strangler-backend: facade/adapter, shadow, cutover, rollback, parity, and cost.

Gate rule

strangler-backend code mode cannot start until this skill records:

• Azure target service choices per slice;
• Well-Architected pillar tradeoffs;
• networking/identity/security constraints;
• observability and rollback plan;
• open Azure environment/tooling gaps.

Output template and helper scripts

• Korean template: OUTPUT_TEMPLATE.md
• Azure guidance reference: AZURE_REFERENCE.md
• Report scaffold: scripts/create-azure-architecture.sh <repo-root> <target-name>
• Tooling check: scripts/check-azure-tooling.sh