ops-wellarchitected-review

name	ops-wellarchitected-review
description	AWS Well-Architected Framework 6-pillar infrastructure review with 100-point scoring, severity-rated findings, and AS-IS to TO-BE transformation roadmap
triggers	["well-architected","WAF review","well architected review","pillar review","인프라 진단","아키텍처 리뷰","심층 진단","WAF 점검","인프라 점수"]
model	sonnet
allowed-tools	["Bash","Read","Grep"]

Well-Architected Framework Review Workflow

Systematic 6-pillar infrastructure assessment producing quantified scores, prioritized findings, and a transformation roadmap.

Phase 1: Scope & Context

1. Determine review scope:
   • Full 6-pillar review (default, ~15-20 min)
   • Specific pillar deep dive (user specifies: Cost, Security, Reliability, Performance, OpEx, Sustainability)
2. Identify target environment:
   • AWS account and region
   • EKS cluster name (if applicable)
   • Key service categories in use
3. Baseline inventory:

# Account and region
aws sts get-caller-identity
aws configure get region

# Service summary
aws ec2 describe-instances --query 'Reservations[].Instances[?State.Name==`running`] | length(@)'
aws rds describe-db-instances --query 'DBInstances | length(@)'
aws eks list-clusters --query 'clusters[]' --output text
aws s3api list-buckets --query 'Buckets | length(@)'
aws lambda list-functions --query 'Functions | length(@)'

Phase 2: Data Collection

Gather data by pillar. Run commands from reference files:

Pillar	Key Data Sources	Reference
Cost Optimization	CE API (cost/usage, RI/SP coverage), idle resource scan	references/pillar-cost-optimization.md
Security	SG audit, encryption check, IAM credential report, public exposure	references/pillar-security-reliability.md Part A
Reliability	VPC/AZ distribution, NAT GW count, Multi-AZ DB, ASG coverage, backup retention	references/pillar-security-reliability.md Part B
Performance	Instance utilization, kubectl top, Lambda memory, DB metrics	references/pillar-performance-opex-sustainability.md Part A
OpEx	CloudWatch alarms, log retention, CloudTrail, IaC detection	references/pillar-performance-opex-sustainability.md Part B
Sustainability	Graviton ratio, gp2/gp3 ratio, Lambda architecture, serverless count	references/pillar-performance-opex-sustainability.md Part C

Missing data handling: If a service is not detected (no EKS clusters, no MSK, etc.), skip that sub-assessment and note: "[Service] not detected — skipping related assessment." Do not fabricate data.

Phase 3: Pillar Assessment

For each pillar, apply scoring criteria from references/waf-scoring-framework.md:

1. Calculate sub-category scores (0-100 scale per sub-category)
2. Apply weights to get pillar score
3. Rate findings by severity:

Severity	Criteria	Examples
Critical	Immediate exploitation/failure risk	Public S3, unencrypted prod DB, single-AZ prod
Warning	Elevated risk, planned remediation	Old access keys, no Multi-AZ dev DB, gp2 volumes
Info	Best practice recommendation	Graviton candidates, lifecycle policy missing

Phase 4: Scoring Synthesis

1. Calculate Infrastructure Health Score (weighted sum of pillar scores):

Score = Σ (pillar_score × pillar_weight)

2. Generate pillar summary table:

Pillar	Score	Weight	Weighted	Status
Operational Excellence	XX	15%	X.X	Good/Fair/Needs Attention
Security	XX	20%	X.X	...
Reliability	XX	20%	X.X	...
Performance Efficiency	XX	15%	X.X	...
Cost Optimization	XX	20%	X.X	...
Sustainability	XX	10%	X.X	...
Total		100%	XX/100	[Rating]

3. Top 5 findings — rank by impact (severity × blast radius)
4. Estimated savings — aggregate from Cost and Performance findings

Phase 5: Roadmap Generation

Generate AS-IS → TO-BE recommendations using the Priority Matrix from references/waf-scoring-framework.md:

Quick Wins (This Week)

#	Finding	AS-IS	TO-BE	Est. Impact	Effort

Short-term (1-3 Months)

#	Finding	AS-IS	TO-BE	Est. Impact	Effort

Medium-term (3-6 Months)

#	Finding	AS-IS	TO-BE	Est. Impact	Effort

Include dependency notes (e.g., "Requires VPC endpoint before NAT GW removal").

Phase 6: Report Delivery

Output the structured WAF report. Match user's language (Korean or English).

============================================================
  AWS Well-Architected Review Report
============================================================
  Date:    YYYY-MM-DD
  Account: XXXXXXXXXXXX
  Region:  ap-northeast-2
  Scope:   Full 6-Pillar Review
============================================================
  Infrastructure Health Score: XX / 100  [Rating]
============================================================

  [Pillar Scores Table]
  [Top 5 Findings]
  [Estimated Savings: Monthly $X, Annual $Y]
  [Quick Wins / Short-term / Medium-term Roadmap]
  [Immediate Action Items]
============================================================

Cross-Agent Delegation

When a pillar scores below 60, recommend delegation to specialist:

Pillar < 60	Delegate To	Action
Cost Optimization	cost-agent	Deep FinOps analysis
Security	iam-agent + ops-security-audit	Full security audit
Reliability (Network)	network-agent + ops-network-diagnosis	Network deep dive
Performance (Compute)	eks-agent	EKS/compute optimization
Performance (Data)	database-agent	Database right-sizing

References

• references/waf-scoring-framework.md — Scoring methodology, weights, criteria, priority matrix
• references/pillar-cost-optimization.md — Cost assessment commands, pricing benchmarks, idle detection
• references/pillar-security-reliability.md — Security scoring, public exposure, encryption, reliability checks
• references/pillar-performance-opex-sustainability.md — Right-sizing, monitoring, Graviton, sustainability metrics