Skip to content

SAA-C03 One-Page Cram Sheet

Well-Architected Pillars

Operational Excellence | Security | Reliability | Performance | Cost Optimization | Sustainability

Storage Decision Tree

  • S3: Object storage, static websites, backup, data archiving
  • EBS: Block storage for EC2, high IOPS, snapshots
  • EFS: Shared file storage, NFS, multi-AZ
  • FSx: High-performance file systems (Windows, Lustre)

Compute Options

  • EC2: Virtual servers, full control, wide instance types
  • Lambda: Serverless, event-driven, 15min max, stateless
  • ECS: Container orchestration, Docker
  • EKS: Managed Kubernetes
  • Fargate: Serverless containers

Database Selection

  • RDS: OLTP, MySQL/PostgreSQL/SQL Server/Oracle, Multi-AZ
  • Aurora: Cloud-native, faster than RDS, auto-scaling
  • DynamoDB: NoSQL, single-digit ms, serverless, DAX for caching
  • Redshift: Data warehouse, OLAP, columnar
  • ElastiCache: In-memory, Redis/Memcached

Networking Essentials

  • VPC: Isolated network, 10.0.0.0/16 default
  • Subnets: Public (route to IGW), Private (route to NAT)
  • Security Groups: Stateful, whitelist, instance-level
  • NACLs: Stateless, subnet-level, allow/deny rules
  • Route Tables: Direct traffic, one per subnet

Security Quick Hits

  • IAM: Users, Groups, Roles, Policies (JSON)
  • Principle of Least Privilege: Minimum required permissions
  • MFA: Multi-factor authentication for privileged access
  • CloudTrail: API logging, governance, compliance
  • Config: Resource compliance monitoring

Load Balancing

  • ALB: HTTP/HTTPS, layer 7, advanced routing
  • NLB: TCP/UDP, layer 4, ultra-high performance
  • CLB: Legacy, layer 4/7, simple use cases
  • Target Groups: Health checks, routing targets

Auto Scaling

  • Scaling Policies: Target tracking, step, simple
  • Metrics: CPU, memory, custom CloudWatch metrics
  • Cooldown: Prevent rapid scaling events
  • Health Checks: EC2 + ELB health checks

High Availability Patterns

  • Multi-AZ: Different availability zones in same region
  • Multi-Region: Geographic distribution, latency reduction
  • Auto Scaling: Horizontal scaling for resilience
  • Load Balancing: Distribute traffic across healthy instances
  • Database: RDS Multi-AZ, Aurora replicas

Cost Optimization

  • Right Sizing: Match instance size to workload
  • Reserved Instances: 1-3 year commitment, up to 75% savings
  • Spot Instances: Unused capacity, up to 90% discount
  • S3 Storage Classes: IA, One Zone-IA, Glacier for archival
  • CloudWatch: Monitor and optimize resource usage

Exam Keywords & Services

High Availability → Multi-AZ, ELB, Auto Scaling Scalability → Auto Scaling Groups, ELB, CloudFront Security → IAM, VPC, Security Groups, encryption Cost → Reserved Instances, Spot, right-sizing Performance → CloudFront, ElastiCache, read replicas Serverless → Lambda, API Gateway, DynamoDB, S3 Monitoring → CloudWatch, CloudTrail, Config Integration → SQS, SNS, SWF, Step Functions