Exam readiness

Tips

Most questions are scenario based
- First rule out answers you know for sure are wrong
- Remaining answers: Understand which one makes the most sense.
No penalty for guessing
Don't over-think it
- If solution seems feasible but highly complicated, it's probably wrong.
You can come back to questions with flag feature
For learning services more, read each service's FAQ
- Cover a lot of the questions asked in the exam
If AWS provides a service/functionality within service => Use it instead of creating a bespoke service.
65 questions will be asked in 130 minutes
- 2 minutes per question

Reliable/resilient storage = EBS, EFS, S3
- For EBS
  - Small, random = SSD
  - Large, sequential = HDD
Durable = S3
Low latency access = EBS
Reproducible = Spot instance, S3 one zone IA
Decoupling = SQS, SNS
- Only once = SQS FIFO or SWF
Multi-tier architecture: ELB, ASG, EC2
High availability/fault tolerant =
- Horizontal scaling, vertical scaling
- ASG that span over many AZ (sent to ELB)
- DynamoDB (HA by default)
- RDS (multi-AZ needs to be enabled)
  - Availability = Sync replication, Scalability = async
Performant storage and databases =
- 💡GP2, IO1 for EBS
- 💡EFS, S3
Caching = ElastiCache, CloudFront, API Gateway, DAX for DynamoDB
Multi-region=Stack Set
Elasticity and scalability = ASG, Lambda, CloudWatch for alarms
CloudWatch
- Enhanced monitoring: Metrics within hyper-visor from RDS
- Detailed monitoring: Send metrics in 1m interval instead of 5.
Secure application tiers = Security groups
Secure data = KMS, Encryption
Cost optimized storage = S3, Pay for what you use, life-cycle rules to move data to cheaper, glacier for long time.
Cost optimized compute = Lambda, pay for what you use, reserve long time instances, Spot Instances for short fault tolerant jobs
Licensing problems = Microsoft + Oracle e.g. BYOB=Dedicated host
Operational excellence = Serverless