AWS SAA-C03 Design Cost-Optimized Architectures Practice Questions

Free SAA-C03 Design Cost-Optimized Architectures practice test questions

8 questions from this domain with answers and explanations - different from the samples on the main AWS SAA-C03 page. Sign up free to practice the full set.

1. Question 1Design Cost-Optimized Architectures

   A data analytics company runs nightly batch processing jobs on a fleet of 20 EC2 instances. Each job takes 3-4 hours to complete and can be restarted from a checkpoint if interrupted. The workload has been running on On-Demand instances, and the CTO wants to reduce compute costs by at least 60%. Which EC2 purchasing option should the solutions architect recommend?

   • AOn-Demand Instances covered by a 1-year Compute Savings Plan that commits to a fixed hourly compute spend regardless of how few hours the nightly jobs actually run
   • BStandard Reserved Instances with a 1-year term and all-upfront payment reserving the full 20-instance fleet around the clock even though the jobs run only at night
   • CSpot Instances with checkpointing and a fleet diversified across instance types and AZsCorrect
   • DDedicated Hosts purchased with a 1-year reservation to guarantee dedicated physical capacity for the nightly batch jobs and to satisfy any per-socket software licensing needs
   ✓ Correct answer: C

   Spot Instances offer up to 90% discount compared to On-Demand pricing, easily meeting the 60% cost reduction goal. Since the workload can be restarted from checkpoints, interruptions are handled gracefully. Using multiple instance types and Availability Zones ensures that if one type or zone has capacity constraints, others are available. This strategy can achieve the 60% cost reduction while maintaining availability for batch processing.

   Why the other options are wrong

   • AA Compute Savings Plan typically yields up to about 30 percent off On-Demand, which falls short of the 60 percent target, and it commits to steady spend that does not suit a job running only a few hours each night.
   • BReserved Instances discount around 40 percent for a 1-year all-upfront term and bill for 24 hours a day, so they are wasteful for a workload that runs only 3 to 4 hours nightly and cannot reach the 60 percent goal.
   • DDedicated Hosts are the most expensive option, aimed at licensing and isolation needs, so they would raise costs rather than cut them by 60 percent for an interruption-tolerant batch job.
2. Question 2Design Cost-Optimized Architectures

   TerraWatt Energy runs batch data processing jobs that can tolerate interruptions and typically take 2-4 hours to complete. The jobs are not time-sensitive and can be restarted if interrupted. Which EC2 purchasing option provides the greatest cost savings?

   • ADedicated Hosts.
   • BOn-Demand Instances.
   • CReserved Instances with a 1-year term.
   • DSpot Instances.Correct
   ✓ Correct answer: D

   Spot Instances use spare EC2 capacity and can be priced up to 90% below On-Demand rates. Because TerraWatt Energy's batch jobs tolerate interruptions and can be restarted, Spot is the ideal purchasing option. The 2-4 hour duration per job means most jobs complete before an interruption occurs, and a restart strategy (such as checkpointing or using Spot-aware job schedulers) handles the rare reclaim events without data loss.

   Why the other options are wrong

   • ADedicated Hosts provide a physical server dedicated to your use for software licensing compliance or regulatory requirements; they are among the most expensive EC2 options and offer no advantage for this interruptible batch workload.
   • BOn-Demand Instances charge the full on-demand rate for every hour used with no discount; while flexible, they provide no cost savings compared to the deeply discounted Spot option available for this workload.
   • CReserved Instances with a 1-year term commit you to paying for capacity whether used or not, typically saving 30-40% versus On-Demand, which is far less than the up-to-90% Spot discount available for restartable jobs.
3. Question 3Design Cost-Optimized Architectures

   Redwood Analytics runs a development environment with 15 EC2 instances that developers use during business hours (8 AM to 6 PM) on weekdays only. The instances are currently running 24/7. Which approach provides the MOST cost savings with minimal operational effort?

   • AConvert all development instances to Spot Instances.
   • BPurchase Reserved Instances for the development workload.
   • CUse AWS Instance Scheduler to start and stop instances on a schedule.Correct
   • DAsk each developer to manually stop their own instances at the end of every day.
   ✓ Correct answer: C

   AWS Instance Scheduler automates EC2 instance start and stop operations based on custom schedules, eliminating idle costs during non-business hours. For development environments used only during business hours, automated scheduling provides superior cost reduction compared to manual approaches. This solution offers the most cost savings with minimal operational effort.

   Why the other options are wrong

   • ASpot Instances suit fault-tolerant workloads and can be reclaimed with two minutes' notice, disrupting developers who need availability during business hours.
   • BReserved Instances discount continuous 24/7 use and require a 1- or 3-year commitment, so they do not suit a part-time weekday-only development environment.
   • DRelying on developers to manually stop instances is error-prone and inconsistent, leaving instances running overnight and delivering less savings than automated scheduling.
4. Question 4Design Cost-Optimized Architectures

   A company uses Amazon RDS for PostgreSQL with a Multi-AZ deployment. The database is running a db.r6g.2xlarge instance. CloudWatch metrics show that the average CPU utilization is 15% and memory usage is 20%. The company wants to reduce costs. What should the architect recommend?

   • ASwitch to a db.r6g.4xlarge instance to get a better price-performance ratio.
   • BRight-size the instance down to a db.r6g.large based on the metrics.Correct
   • CDisable the Multi-AZ deployment to roughly halve the instance cost.
   • DMigrate the relational workload from RDS PostgreSQL to Amazon DynamoDB.
   ✓ Correct answer: B

   Right-sizing analyzes actual resource utilization and adjusts the instance class to match. With average CPU at 15% and memory at 20%, the db.r6g.2xlarge is heavily over-provisioned, so moving to a smaller db.r6g.large in the same family reduces cost while comfortably serving the workload and keeping the Multi-AZ high-availability configuration intact.

   Why the other options are wrong

   • AMoving to a 4xlarge doubles the size and cost of an already heavily over-provisioned instance, which increases spend rather than reducing it.
   • CTurning off Multi-AZ removes the high-availability failover this production database needs; cost should be cut by right-sizing, not by sacrificing resilience.
   • DMigrating to DynamoDB requires re-architecting the relational schema and queries and is unwarranted when simply resizing the over-provisioned instance saves cost.
5. Question 5Design Cost-Optimized Architectures

   A company sends 10 million push notifications daily. They are currently using EC2 instances running custom notification software. Which managed AWS service can replace this setup and reduce costs?

   • AAWS Step Functions workflows
   • BAmazon Kinesis Data Streams
   • CAmazon SQS message queues
   • DAmazon SNSCorrect
   ✓ Correct answer: D

   Amazon Simple Notification Service (SNS) is a fully managed publish/subscribe service built to fan out push notifications, including mobile push, to large numbers of subscribers. Replacing the custom EC2-based notification software with SNS removes the servers to manage and pays only per message, reducing cost for a 10 million per day volume.

   Why the other options are wrong

   • AStep Functions orchestrates multi-step workflows; it coordinates tasks and is not a push-notification delivery service.
   • BKinesis Data Streams ingests and processes streaming data records; it does not deliver push notifications to endpoints.
   • CSQS is a message queue for decoupling components with pull-based consumers; it does not push notifications to devices or subscribers.
6. Question 6Design Cost-Optimized Architectures

   A startup's serverless application uses AWS Lambda functions that each allocate 1,024 MB of memory but, according to monitoring, never use more than 200 MB and finish quickly. The team wants to reduce Lambda cost. Which approach is most effective?

   • ARight-size the memory allocation downward using AWS Lambda Power Tuning insightsCorrect
   • BSwitch the functions to arm64 Graviton architecture to lower the per-GB-second price
   • CEnable SnapStart on the functions to shorten initialization and billed duration
   • DConfigure a longer batch window so more events are processed per invocation
   ✓ Correct answer: A

   Lambda bills on GB-seconds (allocated memory multiplied by execution time), so cutting an over-provisioned 1,024 MB allocation that uses only 200 MB lowers cost directly. AWS Lambda Power Tuning helps find the optimal memory/cost balance, since reducing memory too far can lengthen duration.

   Why the other options are wrong

   • BMoving to Graviton offers a modest per-GB-second price reduction but leaves the function heavily over-provisioned at 1,024 MB; reducing memory to match the 200 MB actually used yields far greater savings.
   • CSnapStart reduces cold-start latency for Java functions and does not lower cost for functions that already finish quickly; the over-allocated memory remains the primary cost driver.
   • DA longer batch window aggregates events for event-source mappings but does not reduce the cost of paying for memory the function never uses, so the over-provisioning persists.
7. Question 7Design Cost-Optimized Architectures

   A company runs analytics queries directly against 100 TB of CSV and JSON files stored in Amazon S3. They want to query the data with standard SQL while paying only for the data scanned, with no clusters to manage. Which service is most cost-effective?

   • AAmazon Redshift provisioned cluster
   • BAmazon AthenaCorrect
   • CAmazon EMR with Apache Hive
   • DAmazon RDS for PostgreSQL
   ✓ Correct answer: B

   Athena is serverless and bills per terabyte scanned, so for occasional or ad-hoc SQL over data already in S3 there is no idle cluster cost. Compressing or converting data to columnar formats further reduces the scanned-bytes cost.

   Why the other options are wrong

   • AA Redshift provisioned cluster runs and bills continuously and requires loading data, adding cost not justified for query-in-place use.
   • CEMR with Hive provisions a cluster that bills while running and adds management overhead versus a serverless query engine.
   • DRDS for PostgreSQL is an OLTP database requiring ingestion and is not designed to query raw files in S3 cost-effectively.
8. Question 8Design Cost-Optimized Architectures

   A company wants its finance and engineering teams to allocate AWS costs to specific projects and cost centers across many resources. Which feature best enables accurate cost allocation and reporting?

   • BCost allocation tags activated in the Billing console and used in Cost ExplorerCorrect
   • AAWS Cost and Usage Report queried in Athena grouped by linked account
   • CAWS Organizations consolidated billing grouped by organizational unit
   • DAWS Resource Groups used to organize resources by application
   ✓ Correct answer: B

   By tagging resources with keys like Project or CostCenter and activating those tags for billing, the company can break down costs along business dimensions. Cost Explorer and the CUR then filter and group by these tags for chargeback and analysis.

   Why the other options are wrong

   • AThe Cost and Usage Report gives detailed billing line items, but without cost allocation tags it cannot attribute spend to specific projects and cost centers spanning many resources.
   • CConsolidated billing rolls up charges per account or organizational unit, which does not map to arbitrary projects and cost centers that share resources within accounts.
   • DResource Groups help organize and act on resources operationally but do not surface cost data in billing or Cost Explorer, so they do not enable cost allocation reporting.

How Design Cost-Optimized Architectures is tested

This domain holds 138 of the 675 questions in the SAA-C03 bank, about 20%. The mix is 118 single-answer multiple choice, 19 multiple-response and 1 yes/no scenario, so it is worth practising the formats as well as the content.

Once you have a few attempts recorded, CertGrid scores every domain separately and points you at the weakest one, so you can drill Design Cost-Optimized Architectures on its own rather than re-running full-length mocks.

Other SAA-C03 exam domains

• Design Secure Architectures202 questions
• Design Resilient Architectures176 questions
• Design High-Performing Architectures159 questions
• All SAA-C03 practice questions675 total
• Design Cost-Optimized Architectures study notesKey concepts
• AWS practice examsAll AWS

SAA-C03 Design Cost-Optimized Architectures FAQ