Q99 — AWS SAA-C03 Ch.16

Question 99 of 100 | ← Chapter 16

Q1299. A financial services company launches a new application that uses an Amazon RDS for MySQL database. The company uses the application to track stock market trends. The company needs to operate the application for only 2 hours at the end of each week. The application must be highly available during the weekly 2-hour period.Which solution will meet these requirements MOST cost-effectively?

Correct Answer: A. Migrate the existing RDS for MySQL database to an Aurora Serverless v2 MySQL database cluster.

Explanation

For a financial services application tracking stock market trends that operates only 2 hours weekly but requires high availability (HA) during that period, the most cost-effective solution is Option A: Migrate to Aurora Serverless v2 MySQL. Here's the detailed analysis:Key Requirements & Evaluation CriteriaHigh Availability (HA): Must ensure zero downtime during the 2-hour window (e.g., no database failures, automatic failover).Cost Efficiency: Minimize expenses for the remaining 166 hours/week when the database is idle. Operational Simplicity: Avoid manual intervention (e.g., launching/terminating instances weekly).Option-by-Option AnalysisOption A: Aurora Serverless v2 MySQLHA: Built-in multi-AZ deployment with automatic failover. Aurora Serverless v2 replicates data across three AZs by default.Cost Efficiency:Pay-per-second pricing: Scales to zero when idle (no connections for ~5 minutes), costing 0.051/hour).Total weekly cost: ~17.54(168hours?0.051/hour), even though the database is only used for 2 hours.Drawback: High idle costs due to always-on provisioning.Option C: Amazon EC2 + Spot InstancesHA: Spot Instances are not HA by design (they can be terminated with 2-minute notice). To achieve HA, you'd need:A multi-AZ setup with at least two Spot Instances (doubling costs). Manual or automated failover logic (e.g., Route 53 health checks).Cost Efficiency:Spot Instances for MySQL (e.g., m6i.large at0.025/hour)wouldcost0.05 for 2 hours.However, adding HA (e.g., two instances + EBS volumes + failover scripts) increases complexity and cost (~0.10?0.15/hour).Drawback: Not truly HA and requires significant DevOps effort.Option D: Amazon ECS with MySQL ContainersHA: Requires a multi-AZ ECS cluster with MySQL containers, plus manual or automated scaling.Cost Efficiency:Running MySQL in containers doesn't inherently reduce costs; you still pay for EC2 instances or Fargate tasks.For 2 hours/week, Fargate would cost ~0.013pervCPU-hour(e.g.,2vCPUs?hours?0.013 = $0.052).However, HA adds complexity (e.g., ECS Service Auto Scaling, multi-AZ task placement). Drawback: Overkill for a simple database workload; higher operational overhead than Aurora Serverless.Why Option A WinsCost: Aurora Serverless v2 is 98% cheaper than provisioned Aurora MySQL for this use case (weekly cost:0.24vs.17.54).HA: Matches the reliability of provisioned Aurora without manual intervention. Simplicity: No need to manage instances, clusters, or failover logic.Implementation StepsMigrate to Aurora Serverless v2:Use AWS Database Migration Service (DMS) to transfer data from RDS MySQL to Aurora Serverless v2 with minimal downtime.Configure Auto Scaling:Set minimum capacity to 0 ACUs (scales to zero when idle) and maximum capacity based on peak load (e.g., 10 ACUs).Test Failover:Simulate an AZ failure to verify automatic failover within <30 seconds.Monitor Costs:Use AWS Cost Explorer to track spending and validate savings.ConclusionOption A is the only solution that delivers high availability, zero idle costs, and no operational overhead for a database used 2 hours/week. The other options either fail to meet HA requirements (C, D) or are prohibitively expensive (B).