AWS Certified Solutions Architect Associate · Free Practice Question Hard
Question 55
A financial data processing company runs a workload on Amazon EC2 instances that fetch and process real-time transaction batches from an Amazon SQS queue. The application needs to scale based on unpredictable message volume, which fluctuates significantly throughout the day. The system must process messages with minimal delay and no downtime, even during peak spikes. The company is seeking a solution that balances cost-efficiency with availability and elasticity.
Which EC2 purchasing strategy best meets these requirements in the most cost-effective manner?
-
A
Use EC2 Reserved Instances for the baseline workload and configure EC2 Auto Scaling to launch On-Demand Instances for all traffic spikes
-
B
Use Reserved Instances for the baseline level of traffic and configure EC2 Auto Scaling with Spot Instances to handle spikes in message volume
-
C
Use EC2 Spot Instances exclusively with Auto Scaling enabled to match message volume fluctuations and save on compute costs
-
D
Purchase EC2 Reserved Instances to match peak capacity and assign all message processing tasks to these instances regardless of load variations
Reveal correct answer
Correct answer: B
Explanation
Correct option:
Use Reserved Instances for the baseline level of traffic and configure EC2 Auto Scaling with Spot Instances to handle spikes in message volume
This blended approach is ideal for workloads that are partially predictable and partially variable. Reserved Instances ensure cost-efficient baseline capacity for steady message ingestion, while Spot Instances are used to handle unpredictable surges in traffic, significantly reducing costs compared to On-Demand. If Spot capacity is unavailable, Auto Scaling can be configured with fallback instance types or prioritized mixed instance policies. This model maximizes cost savings without compromising on availability.
Incorrect options:
Purchase EC2 Reserved Instances to match peak capacity and assign all message processing tasks to these instances regardless of load variations - While Reserved Instances (RIs) offer significant cost savings compared to On-Demand pricing, they are best suited for predictable, steady-state workloads. Using RIs to provision maximum capacity at all times, regardless of actual load, results in over-provisioning and wasted cost during low-traffic periods. This approach lacks elasticity and is not suitable for workloads with unpredictable and bursty traffic patterns.
Use EC2 Spot Instances exclusively with Auto Scaling enabled to match message volume fluctuations and save on compute costs - Spot Instances provide deep cost savings, but they can be interrupted by AWS with only two minutes of warning if capacity is reclaimed. Using them exclusively for a mission-critical workload with real-time, parallel processing introduces risk of downtime and message loss, especially if traffic spikes coincide with Spot interruptions. For fault-tolerant or batch jobs, this could work—but not for always-on processing with strict uptime requirements.
Use EC2 Reserved Instances for the baseline workload and configure EC2 Auto Scaling to launch On-Demand Instances for all traffic spikes - While combining Reserved Instances for baseline capacity with On-Demand Instances for additional traffic is a functional and reliable solution, it is not the most cost-effective choice for a workload with unpredictable, intermittent spikes like real-time message processing. On-Demand Instances are priced at a premium and can lead to significantly higher costs when scaling frequently. A blended model using Reserved + Spot Instances would better balance availability and cost.
References:
https://docs.aws.amazon.com/autoscaling/ec2/userguide/ec2-auto-scaling-mixed-instances-groups.html
https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-spot-instances.html
https://docs.aws.amazon.com/savingsplans/latest/userguide/what-is-savings-plans.html
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
