Domain 3 of 4 · Chapter 3 of 5

Improving performance

Unlock the complete study guide + 1,040 practice questions across 16 full exams.

Bundled into the existing AWS Certified Solutions Architect - Professional premium course — no separate purchase.

Included in this chapter:

  • The measure-tune-verify loop for live workloads
  • Rightsizing existing compute, storage, and databases
  • Diagnosing and tuning database performance
  • Scaling, edge acceleration, and platform migration
  • Exam patterns: reading an improve-performance question

Performance-improvement tooling for existing AWS workloads: which lens (SAP-C02)

DimensionAWS Compute OptimizerAWS Trusted Advisor (Performance)Amazon DevOps GuruRDS Performance Insights
Primary question it answersIs this resource the right size?Are we following performance best practices?What anomaly just broke (or is about to)?Why is this database slow right now?
MethodUtilization analysis + rightsizing recommendationsRule-based best-practice checksMachine-learning anomaly detection + insightsDB load (Average Active Sessions) vs Max vCPU
Scope of resourcesEC2, ASG, EBS, Lambda, ECS-on-Fargate, Aurora/RDSEC2 utilization, EBS, service quotasApp-wide resources (DevOps Guru for RDS adds DB depth)One RDS/Aurora DB instance
Data window14 days default; 93 days with enhanced metricsPeriodic checks (auto-refresh on Business/Enterprise)Continuous learned baselineReal-time + historical retention
Access requirementFree; opt-in per account/RegionFull checks need Business/Enterprise SupportPaid; opt-in coverage per account/RegionEnable on the DB instance (migrating to Database Insights)
Best forRightsizing a fleet from real dataQuick portfolio-wide hygiene sweepCatching regressions without static thresholdsDiagnosing relational query / wait bottlenecks

Decision tree

Measure first: where is the bottleneck? Request latency Distributed / many downstream calls? Database slow Relational DB load or read storm? Sizing / global Wrong-sized fleet or global latency? Yes AWS X-Ray trace map (find slow segment) No CloudWatch metrics + anomaly detection Diagnose Performance Insights + DevOps Guru for RDS Offload reads Read replicas + ElastiCache; DAX for DynamoDB hot keys Sizing AWS Compute Optimizer (14d / 93d rightsizing) Global CloudFront (cacheable) / Global Accelerator Always: change one variable, then re-test under load — measure the gain, don't assume it.
Cheat sheet

Cheat sheet

  • Measure before you scale
  • CloudWatch anomaly detection beats static thresholds for seasonal load
  • X-Ray trace map locates the slow downstream call
  • Compute Optimizer rightsizes from 14 days of utilization
  • Enhanced infrastructure metrics extend the lookback to 93 days
  • Rightsizing is a performance lever, not only a cost lever
  • Trusted Advisor Performance checks need Business/Enterprise Support
  • Performance Insights: DB load vs the Max vCPU line
  • Performance Insights slices load by waits, SQL, hosts, users
  • DevOps Guru for RDS adds ML over Performance Insights
  • DevOps Guru gives reactive and proactive insights
  • Read replicas offload reads from the primary
  • ElastiCache fronts the database for repeat reads
  • DAX accelerates only eventually-consistent DynamoDB reads
  • DynamoDB auto scaling uses target tracking 20-90%
  • Convert DynamoDB to on-demand for unpredictable traffic
  • CloudFront for cacheable, Global Accelerator for non-cacheable
  • Graviton gives up to 20% lower cost (and up to 60% less energy)
  • gp3 provisions IOPS and throughput independently of size
  • Validate every change with load testing
  • Origin Shield adds a caching layer in the Region nearest the origin
  • Global Accelerator: anycast IPs with traffic dials, weights, and fail-open
  • Pick the right placement group; ENA Express boosts single-flow throughput
  • Scale reads with Multi-AZ DB clusters, read replicas, or Aurora readers

Unlock with Premium — includes all practice exams and the complete study guide.

References

  1. Using Amazon CloudWatch metrics
  2. CloudWatch anomaly detection
  3. What is AWS X-Ray?
  4. What is AWS Compute Optimizer?
  5. Compute Optimizer enhanced infrastructure metrics
  6. AWS Trusted Advisor
  7. Monitoring DB load with Performance Insights
  8. Working with DevOps Guru for RDS
  9. What is Amazon DevOps Guru?
  10. CloudWatch Database Insights for RDS
  11. Amazon RDS read replicas
  12. Amazon ElastiCache: What is ElastiCache?
  13. DynamoDB Accelerator (DAX)
  14. Amazon CloudFront developer guide
  15. What is AWS Global Accelerator?
  16. DynamoDB auto scaling
  17. AWS Graviton processors
  18. Load testing your workload (Performance Efficiency pillar) Well-Architected
  19. Analyzing metrics with the Performance Insights dashboard
  20. Replication with Amazon Aurora
  21. Amazon EC2 Auto Scaling groups
  22. Amazon EBS General Purpose SSD (gp3) volumes