The six pillars of the Well-Architected Framework

• I. Operational Excellence — Running and monitoring systems to deliver business value and continually improving operations. Key practices: infrastructure as code, annotated documentation, frequent small reversible changes, regular game days, post-mortems.
• II. Security — Protecting information, systems, and assets. Key practices: identity foundation (least privilege), traceability (CloudTrail), security at all layers, automated security best practices, protect data in transit and at rest, keep people away from data, prepare for security events.
• III. Reliability — Ensuring a workload performs its intended function correctly and consistently. Key practices: test recovery procedures, automatically recover from failure, scale horizontally, stop guessing capacity, manage change through automation.
• IV. Performance Efficiency — Using computing resources efficiently. Key practices: democratize advanced technologies (use managed services), go global in minutes, use serverless architectures, experiment more often, consider mechanical sympathy (use the right tool for the job).
• V. Cost Optimization — Avoiding unnecessary costs. Key practices: implement cloud financial management, adopt a consumption model (pay for what you use), measure overall efficiency, stop spending on undifferentiated heavy lifting, analyze and attribute expenditure.
• VI. Sustainability — (Added 2021) Minimizing the environmental impacts of running cloud workloads. Key practices: understand your impact, establish sustainability goals, maximize utilization, use managed services, use higher-level managed services that spread load efficiently, reduce downstream impact.

Key reliability patterns with AWS services

• Automatic recovery (circuit breaker) — Amazon EC2 auto recovery, ECS health checks, ALB target group health checks — automatically replace unhealthy instances without human intervention.
• Test recovery procedures — Run game days: simulate AZ failures, instance terminations, database failovers. The Well-Architected Framework says "never guess your RTO — measure it." Most teams discover their actual recovery time is 10× their assumed RTO during game days.
• Horizontal scaling — Auto Scaling Groups with EC2, ECS Service autoscaling, DynamoDB on-demand — add capacity as load increases, remove it when load drops. Never a single point of failure.
• Manage change with automation — CloudFormation, CDK, or Terraform for all infrastructure changes. No manual console changes in production. Every change is code-reviewed, tested in staging, and deployed with rollback capability.