agentleFS
Sign inStart free
kayamankayaman/skills/azure-well-architected/SKILL.md

azure-well-architected

Skill3 starsChanged 8 months agoCC0-1.0View on GitHub

Enforces Azure Well-Architected Framework best practices across all five pillars (Reliability, Security, Cost Optimization, Operational Excellence, Performance Efficiency). Use when designing Azure or application architecture, writing infrastructure code (Bicep, ARM, Terraform), reviewing pull requests for cloud workloads, or making technology selection decisions. Provides both Azure-specific guidance and cloud-agnostic application and solutions architecture principles.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

38 rules. About 0 are common, 37 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈37

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Treat infrastructure as code — define all resources in version-controlled templates

    line 26
  2. Distinctive:

    Prefer managed services over self-managed to reduce operational burden

    line 27
  3. Distinctive:

    Keep it simple — avoid overengineering.

    line 33
  4. Distinctive:

    Focus: Ensure the workload meets uptime and recovery targets by building redundancy and resiliency.

    line 39
  5. Distinctive:

    Keep it simple — avoid overengineering.

    line 47
  6. Distinctive:

    MUST define and document RTO/RPO targets.

    line 51
  7. Distinctive:

    MUST design for multi-zone or multi-region deployments for stateful services in production

    line 52
  8. Distinctive:

    MUST implement health checks and automatic replacement of unhealthy instances

    line 53
  9. Distinctive:

    MUST use decoupling (queues, event buses) to absorb load spikes and isolate failures

    line 54
  10. Distinctive:

    SHOULD use circuit breakers and retry logic with exponential backoff for service calls

    line 55
  11. Distinctive:

    SHOULD simulate failures in pre-production and production.

    line 56
  12. Distinctive:

    MUST enforce least-privilege access.

    line 74
  13. Distinctive:

    MUST enable encryption at rest and in transit for all data stores and communication channels

    line 75
  14. Distinctive:

    MUST classify data by sensitivity and apply appropriate protection per classification level

    line 76
  15. Distinctive:

    MUST maintain an audit trail of access and modification activities

    line 77
  16. Distinctive:

    MUST use Azure Key Vault or equivalent for secrets.

    line 78
  17. Distinctive:

    SHOULD align with Microsoft Zero Trust (verify explicitly, least privilege, assume breach)

    line 79
  18. Distinctive:

    SHOULD define and maintain an incident response plan.

    line 80
  19. Very specific:

    MUST tag all Azure resources with at minimum project, environment, and owner

    line 98
  20. Distinctive:

    MUST set budget alerts and cost thresholds for every workload and subscription

    line 99
  21. Distinctive:

    MUST establish a cost model and baseline.

    line 100
  22. Distinctive:

    SHOULD use Reserved Instances or Savings Plans for predictable, sustained workloads

    line 101
  23. Distinctive:

    SHOULD use Spot VMs for fault-tolerant, batch, or stateless workloads where applicable

    line 102
  24. Distinctive:

    MUST decommission underutilized, unused, or obsolete resources.

    line 103
  25. Distinctive:

    MUST implement structured logging, distributed tracing, and dashboards for every workload

    line 121
  26. Distinctive:

    MUST use automated pipelines for all deployments across all environments

    line 122
  27. Distinctive:

    MUST define operational KPIs tied to business outcomes, not just technical metrics

    line 123
  28. Distinctive:

    MUST use Infrastructure as Code (Bicep, ARM, Terraform) for all infrastructure

    line 124
  29. Distinctive:

    SHOULD prefer small, incremental updates over large releases

    line 125
  30. Distinctive:

    SHOULD use deployment patterns that allow progressive exposure (blue-green, canary, slots)

    line 126
  31. Distinctive:

    SHOULD hold post-incident reviews and feed learnings back into procedures

    line 127
  32. Distinctive:

    MUST define performance targets for critical flows.

    line 144
  33. Distinctive:

    MUST configure auto-scaling for compute resources to match demand

    line 145
  34. Distinctive:

    MUST conduct performance testing before production and after significant changes

    line 146
  35. Distinctive:

    MUST monitor end-to-end performance.

    line 147
  36. Distinctive:

    SHOULD right-size resources across the stack.

    line 148
  37. Distinctive:

    SHOULD use caching and CDN where latency reduction is needed

    line 149
  38. Distinctive:

    Use this checklist when proposing or reviewing Azure or application architecture.

    line 155

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Azure Well-Architected Framework

Reference: https://learn.microsoft.com/en-us/azure/well-architected/

When to Apply

Apply this skill whenever:

  • Proposing or reviewing Azure or application architecture decisions
  • Writing or reviewing IaC (Bicep, ARM templates, Terraform)
  • Reviewing pull requests that touch cloud infrastructure or Azure service integrations
  • Selecting Azure services or making technology trade-off decisions
  • Evaluating cost, performance, security, or reliability implications of a change
  • Designing solutions that require general application or solutions architecture guidance

General Architecture Principles

These cloud-agnostic principles apply across all pillars:

  • Design for business requirements — align architecture with documented goals, constraints, and success criteria; avoid guesswork
  • Treat infrastructure as code — define all resources in version-controlled templates
  • Prefer managed services over self-managed to reduce operational burden
  • Design for failure — assume components will fail; build resilience and fault isolation
  • Adopt zero-trust security — verify explicitly, least privilege, assume breach
  • Automate operations — deployments, scaling, remediation, and testing
  • Make frequent, small, reversible changes; avoid large, risky releases
  • Collect and act on observability data (metrics, logs, traces) at every layer
  • Keep it simple — avoid overengineering; add components only when they achieve target business value

Pillar 1: Reliability

Focus: Ensure the workload meets uptime and recovery targets by building redundancy and resiliency.

Design Principles

  • Design for business requirements — define clear, achievable goals negotiated with stakeholders
  • Design for resilience — expect failures; build fault-tolerant systems that degrade gracefully
  • Design for recovery — have tested, documented recovery plans aligned with RTO/RPO
  • Design for operations — shift left; test failures early; build observable systems
  • Keep it simple — avoid overengineering; minimize critical path components

Rules

  • MUST define and document RTO/RPO targets; MUST test recovery procedures regularly
  • MUST design for multi-zone or multi-region deployments for stateful services in production
  • MUST implement health checks and automatic replacement of unhealthy instances
  • MUST use decoupling (queues, event buses) to absorb load spikes and isolate failures
  • SHOULD use circuit breakers and retry logic with exponential backoff for service calls
  • SHOULD simulate failures in pre-production and production; run chaos experiments

Pillar 2: Security

Focus: Protect the workload from attacks; maintain confidentiality, integrity, and availability (CIA triad).

Design Principles

  • Plan security readiness — adopt security practices in design and operations with minimal friction
  • Protect confidentiality — access controls, data classification, encryption, guard against exfiltration
  • Protect integrity — prevent tampering of data, code, and operations; verify with cryptography
  • Protect availability — prevent downtime from security incidents; balance security and access
  • Sustain and evolve — continuously improve; assume attackers evolve; measure and enforce posture

Rules

  • MUST enforce least-privilege access; MUST NOT use wildcard or overly broad permissions in production
  • MUST enable encryption at rest and in transit for all data stores and communication channels
  • MUST classify data by sensitivity and apply appropriate protection per classification level
  • MUST maintain an audit trail of access and modification activities
  • MUST use Azure Key Vault or equivalent for secrets; MUST NOT hardcode credentials
  • SHOULD align with Microsoft Zero Trust (verify explicitly, least privilege, assume breach)
  • SHOULD define and maintain an incident response plan; SHOULD conduct threat modeling

Pillar 3: Cost Optimization

Focus: Maximize return on investment; optimize usage, rate, and allocation within budget.

Design Principles

  • Develop cost-management discipline — budgets, accountability, reporting, cost tracking
  • Design with cost-efficiency mindset — spend only on what is needed; establish cost baseline
  • Design for usage optimization — maximize resource utilization; right-size and scale dynamically
  • Design for rate optimization — leverage commitments, reservations, consumption models
  • Monitor and optimize over time — right-size as workload evolves; decommission unused resources

Rules

  • MUST tag all Azure resources with at minimum project, environment, and owner
  • MUST set budget alerts and cost thresholds for every workload and subscription
  • MUST establish a cost model and baseline; MUST track expense at workload level
  • SHOULD use Reserved Instances or Savings Plans for predictable, sustained workloads
  • SHOULD use Spot VMs for fault-tolerant, batch, or stateless workloads where applicable
  • MUST decommission underutilized, unused, or obsolete resources; delete unnecessary data

Pillar 4: Operational Excellence

Focus: Build and run workloads correctly; minimize process variance and human error.

Design Principles

  • Embrace DevOps culture — collaboration, shared responsibility, continuous improvement
  • Establish development standards — quality gates, source control, documentation, progress tracking
  • Evolve operations with observability — visibility, insight, data-driven decisions; health modeling
  • Automate for efficiency — replace repetitive manual tasks with automation
  • Adopt safe deployment practices — guardrails, incremental updates, rollback capability

Rules

  • MUST implement structured logging, distributed tracing, and dashboards for every workload
  • MUST use automated pipelines for all deployments across all environments
  • MUST define operational KPIs tied to business outcomes, not just technical metrics
  • MUST use Infrastructure as Code (Bicep, ARM, Terraform) for all infrastructure
  • SHOULD prefer small, incremental updates over large releases
  • SHOULD use deployment patterns that allow progressive exposure (blue-green, canary, slots)
  • SHOULD hold post-incident reviews and feed learnings back into procedures

Pillar 5: Performance Efficiency

Focus: Adapt to changing demands; scale effectively and sustain performance targets.

Design Principles

  • Negotiate realistic performance targets — align with business; define benchmarks and thresholds
  • Design for capacity requirements — provide enough supply for anticipated demand
  • Achieve and sustain performance — protect against degradation; test and monitor continuously
  • Optimize for long-term improvement — reassess targets; adopt new technologies as appropriate

Rules

  • MUST define performance targets for critical flows; MUST align with business stakeholders
  • MUST configure auto-scaling for compute resources to match demand
  • MUST conduct performance testing before production and after significant changes
  • MUST monitor end-to-end performance; MUST set alerts on regressions
  • SHOULD right-size resources across the stack; SHOULD match services to access patterns
  • SHOULD use caching and CDN where latency reduction is needed

Architecture Review Checklist

Use this checklist when proposing or reviewing Azure or application architecture.

Reliability

  • RTO and RPO are defined; recovery procedures are tested and documented
  • Workload is deployed across multiple zones or regions for production
  • Health checks and auto-healing are configured
  • Components are decoupled using queues or event buses
  • Failure scenarios have been simulated (chaos, game days)

Security

  • All access follows least privilege; no wildcard permissions in production
  • All data at rest and in transit is encrypted
  • Secrets are managed via Key Vault or equivalent, not hardcoded
  • Data is classified; protection matches classification level
  • Incident response plan exists; threat modeling has been performed

Cost Optimization

  • All resources are tagged with project, environment, owner
  • Budget alerts and cost thresholds are configured
  • Cost model and baseline are established
  • Idle or unused resources have been removed or scheduled
  • Reservations or commitments evaluated for predictable workloads

Operational Excellence

  • Observability is defined: metrics, structured logs, and distributed traces
  • Deployments are automated via CI/CD with rollback capability
  • Infrastructure is defined as code (Bicep, ARM, or Terraform)
  • Operational runbooks exist for known failure scenarios
  • KPIs tied to business outcomes are identified and monitored

Performance Efficiency

  • Performance targets are defined for critical flows
  • Auto-scaling is configured for compute resources
  • Storage and database types match access patterns
  • Load and performance testing has been performed
  • Caching is applied where latency reduction is needed

Trade-off Guidance

Common tensions between pillars and how to resolve them:

TensionGuidance
Security vs. PerformancePrefer TLS termination at load balancer; use regional endpoints to minimize latency; apply security controls without blocking legitimate users
Cost vs. ReliabilityUse multi-zone for stateful tiers; accept single-zone for stateless compute only in non-critical environments
Cost vs. PerformanceUse reserved capacity for baseline; consumption-based for burst; cache aggressively
Operational simplicity vs. ReliabilityPrefer managed services even at higher unit cost; reduced operational burden outweighs marginal cost difference
Security vs. ReliabilitySecurity controls can introduce points of failure; design compensating controls and avoid single points of failure in security layers

Additional Resources

For detailed per-pillar guidance and Azure-specific services, see reference.md.

Official documentation:

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free