well-architected

GCP Well-Architected Framework

Six pillars that every GCP workload should be evaluated against. Referenced in every HLD, security review, and operational readiness check.

The Six Pillars

Pillar	Core Question	Key Practices
Operational Excellence	Can we build, run, and evolve the system reliably?	IaC, CI/CD, runbooks, monitoring-as-code
Security	Is the system protected against threats?	Least-privilege IAM, encryption, no public exposure by default
Reliability	Will the system recover from failures?	Multi-region, health checks, SLOs, chaos testing
Cost Optimization	Are we spending efficiently?	Right-sizing, committed use, lifecycle policies, budget alerts
Performance	Can the system handle load and scale?	Autoscaling, CDN, caching, regional proximity
Sustainability	Are we minimizing environmental impact?	Carbon-aware regions, efficient resource use

Quick Review Checklist (per design)

Operational Excellence

• Infrastructure as Code (Terraform or Config Connector)
• CI/CD pipeline with automated tests
• Runbook for each primary failure mode
• Alerting tied to SLOs

Security

• Least-privilege IAM throughout
• No public exposure without intent
• Secrets in Secret Manager
• Network deny-by-default

Reliability

• SLO defined (availability + latency)
• Health checks configured
• Multi-zone or multi-region for production
• Backup and recovery tested

Cost Optimization

• Right-sized instances (use GCP recommender)
• Committed use discounts evaluated
• Budget alerts configured
• Lifecycle policies on storage

Performance

• Load tested at target traffic
• Caching layer evaluated
• CDN for static assets
• Autoscaling configured

Sustainability

• Low-carbon region selected where latency allows
• Idle resources scheduled for shutdown

References

• Google Cloud Architecture Framework
• Operational Excellence Pillar
• Security Pillar
• Reliability Pillar
• Cost Optimization Pillar
• Performance Optimization Pillar