generate a high-quality moderately difficult question for the AWS SAA exam with multiple…
- User
generate a high-quality moderately difficult question for the AWS SAA exam with multiple choice answers in the following format
Q1 A fintech start‑up must ingest stock‑market tick data in real time. Each record is < 50 KB, arrives in bursts of up to 30 MB/s, and must be available to downstream consumers within 5 seconds. The company prefers a fully managed solution that minimizes operational overhead and allows replaying data for up to 24 hours. What approach meets the requirements MOST cost‑effectively? A) Amazon MSK with three broker nodes across three AZs B) Amazon Kinesis Data Streams with on‑demand capacity mode and 24‑hour retention C) Amazon SQS standard queue with Lambda pollers D) Amazon EventBridge with an archive rule set to 24 hours A) INCORRECT MSK is always‑on, requires cluster management and is more expensive for bursty workloads. B) CORRECT Fully managed, auto‑scales on demand, supports 24‑hour replay and charges only for used throughput. C) INCORRECT SQS is not optimized for ordered, low‑latency streaming at this scale. D) INCORRECT EventBridge archives events but does not provide millisecond‑latency streaming or high‑throughput ingestion.
Assistant
Q1 An e-commerce company runs its flagship application in the
us-east-1Region. The application consists of EC2 instances for the web and application tiers in an Auto Scaling group behind an Application Load Balancer, and an Amazon RDS for PostgreSQL database. The company needs a disaster recovery solution in theus-west-2Region with a Recovery Time Objective (RTO) of less than 15 minutes and a Recovery Point Objective (RPO) of a few seconds. The failover process must be fully automated. The solution should minimize the cost of the DR environment when it is not active.Which DR strategy meets all these requirements?
A) Deploy a fully scaled-out application stack in
us-west-2and use AWS Global Accelerator to distribute traffic between the two Regions. B) Configure an Amazon RDS for PostgreSQL cross-Region read replica inus-west-2. Inus-west-2, create a scaled-down Auto Scaling group and ALB for the application stack. Use Amazon Route 53 with health checks and failover routing to automatically direct traffic tous-west-2ifus-east-1becomes unavailable. C) Take regular automated snapshots of the EC2 instances' EBS volumes and the RDS database. Copy the snapshots tous-west-2using AWS Backup. In a disaster, use a CloudFormation template to launch a new stack inus-west-2from the latest snapshots. D) Use AWS Elastic Disaster Recovery (DRS) to continuously replicate the EC2 instances and the RDS database server tous-west-2. Use Route 53 failover routing to point to the recovered instances after a failover event. A) INCORRECT This describes a hot standby/multi-site active-active approach. While it provides a near-zero RTO/RPO, it does not meet the requirement to minimize the cost of the idle DR environment, as it maintains a fully scaled, active deployment in the DR region. B) CORRECT This describes a Pilot Light strategy. The RDS cross-Region read replica provides asynchronous replication, meeting the low RPO of a few seconds. Promoting the read replica and scaling up the minimal Auto Scaling group can be done well within the 15-minute RTO. Route 53 health checks and failover routing provide the required automation. Keeping a scaled-down "pilot" infrastructure minimizes costs. C) INCORRECT This is a Backup and Restore strategy. The RTO would likely be hours, not minutes, due to the time required to provision a new stack and restore data from snapshots. The RPO is also much higher than a few seconds, as it is dependent on the snapshot frequency. This fails to meet the RTO and RPO requirements. D) INCORRECT While AWS DRS is a powerful DR tool, using it for a managed service like Amazon RDS is not the standard or most cost-effective pattern. The native RDS cross-Region read replica feature is specifically designed for this purpose, is more tightly integrated, and is the recommended, most efficient solution for replicating RDS databases for disaster recovery.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.