Cost Optimization practice questions

A company has a centralized logging account and multiple application accounts. All VPC Flow Logs are sent to a central S3 bucket in the logging account. The security team needs to analyze the logs using Amazon Athena. The team must ensure queries are cost-effective and return results quickly for recent logs. Which configuration should be used?

Trap 1: Convert the logs to Parquet format using AWS Glue and store them in…

Parquet is good but partitioning is more impactful for recent data.

Trap 2: Use S3 lifecycle policies to transition logs to S3 Glacier after 7…

Glacier is not directly queryable by Athena.

Trap 3: Use Athena with federated query to scan logs directly from the…

Cross-account queries add complexity and latency.

A
Convert the logs to Parquet format using AWS Glue and store them in the same bucket.
Why wrong: Parquet is good but partitioning is more impactful for recent data.
B
Use S3 lifecycle policies to transition logs to S3 Glacier after 7 days and query with Athena.
Why wrong: Glacier is not directly queryable by Athena.
C
Partition the S3 bucket by date (e.g., year/month/day) and use Athena partition projection.
Correct: Partitioning reduces data scanned.
D
Use Athena with federated query to scan logs directly from the application accounts.
Why wrong: Cross-account queries add complexity and latency.

Question 2mediummultiple choice

A company is migrating a monolithic .NET application to AWS. The application uses a SQL Server database with complex stored procedures and tightly coupled components. The migration team wants to minimize refactoring and reduce licensing costs. Which migration strategy should the team use?

Trap 1: Refactor the application to .NET Core and use Amazon RDS for MySQL…

Refactoring is significant and MySQL may not support all stored procedures.

Trap 2: Rearchitect the application to microservices and use Amazon…

Rearchitecting is too disruptive for minimizing refactoring.

Trap 3: Replatform the database to Amazon Aurora with PostgreSQL…

Aurora PostgreSQL would require rewriting stored procedures and may not be compatible.

A
Refactor the application to .NET Core and use Amazon RDS for MySQL to eliminate Windows licensing.
Why wrong: Refactoring is significant and MySQL may not support all stored procedures.
B
Rehost the application on Amazon EC2 and the database on EC2 with SQL Server using Bring Your Own License (BYOL).
This minimizes changes and leverages existing licenses.
C
Rearchitect the application to microservices and use Amazon DynamoDB for the database.
Why wrong: Rearchitecting is too disruptive for minimizing refactoring.
D
Replatform the database to Amazon Aurora with PostgreSQL compatibility to reduce licensing costs.
Why wrong: Aurora PostgreSQL would require rewriting stored procedures and may not be compatible.

Question 3easymultiple choice

A company is using AWS Organizations with consolidated billing. The finance team wants to track costs by business unit. Each business unit has its own AWS account. The team needs a solution that allows them to generate cost reports filtered by business unit without additional overhead. Which action should be taken?

Trap 1: Use Cost Explorer to filter by linked account, which automatically…

Accounts may not map directly to business units.

Trap 2: Use AWS Budgets to create budgets per account and manually…

Manual overhead.

Trap 3: Create a separate payer account for each business unit.

Not needed; consolidated billing is already in place.

A
Use Cost Explorer to filter by linked account, which automatically groups by business unit.
Why wrong: Accounts may not map directly to business units.
B
Use AWS Budgets to create budgets per account and manually aggregate.
Why wrong: Manual overhead.
C
Apply tags to resources and enable cost allocation tags in Billing and Cost Management.
Correct: Tagging enables filtering by business unit.
D
Create a separate payer account for each business unit.
Why wrong: Not needed; consolidated billing is already in place.

Read the full NAT/PAT explanation →

Question 4hardmultiple choice

A company is migrating a monolithic application to microservices on Amazon ECS with Fargate. The application has variable traffic patterns, with high traffic during business hours and low traffic at night. They want to optimize costs while maintaining performance. Which scaling strategy should they implement?

Trap 1: Use step scaling policies based on memory utilization.

Step scaling is more complex, not necessary here.

Trap 2: Use scheduled scaling to increase capacity during business hours.

Scheduled scaling cannot handle unexpected spikes.

Trap 3: Use simple scaling policies based on CPU utilization.

Simple scaling is less responsive.

A
Use target tracking scaling with a schedule to increase minimum capacity during business hours.
This combination handles both patterns.
B
Use step scaling policies based on memory utilization.
Why wrong: Step scaling is more complex, not necessary here.
C
Use scheduled scaling to increase capacity during business hours.
Why wrong: Scheduled scaling cannot handle unexpected spikes.
D
Use simple scaling policies based on CPU utilization.
Why wrong: Simple scaling is less responsive.

Question 5mediummulti select

A company is running a critical application on Amazon EC2 instances in an Auto Scaling group behind an Application Load Balancer. The application needs to process a large batch job that runs once per month and takes 2 hours. The company wants to optimize costs while ensuring the batch job has sufficient capacity. Which THREE steps should a solutions architect recommend?

Trap 1: Use On-Demand Instances only for the batch job to ensure…

On-Demand is costlier than Spot for this workload.

Trap 2: Purchase Reserved Instances for the batch job capacity.

Reserved Instances require a 1- or 3-year commitment, not suitable for short monthly jobs.

A
Create a scheduled scaling policy that increases the desired capacity before the batch job starts.
Scheduled scaling ensures capacity is available when needed.
B
Use a launch template with multiple instance types to increase the chances of obtaining Spot capacity.
Diversifying instance types improves Spot availability.
C
Use On-Demand Instances only for the batch job to ensure availability.
Why wrong: On-Demand is costlier than Spot for this workload.
D
Use Spot Instances in the Auto Scaling group for the batch job.
Spot Instances are cost-effective for fault-tolerant batch workloads.
E
Purchase Reserved Instances for the batch job capacity.
Why wrong: Reserved Instances require a 1- or 3-year commitment, not suitable for short monthly jobs.

Question 6mediummulti select

A company is planning to migrate a legacy application to AWS. The application runs on a single server with a monolithic architecture and uses an Oracle database. The migration team wants to reduce licensing costs and improve scalability. Which TWO strategies should the team consider?

Trap 1: Refactor the application into microservices and use Amazon DynamoDB…

High refactoring effort; not directly reducing licensing costs.

Trap 2: Rehost the application on Amazon EC2 and use Amazon RDS for Oracle…

License-included is more expensive.

Trap 3: Rehost the application on Amazon EC2 and use Amazon RDS for Oracle…

Still uses Oracle licensing; does not reduce costs.

A
Replatform the application to use AWS Elastic Beanstalk and Amazon RDS for PostgreSQL.
Reduces operational overhead and licensing costs.
B
Refactor the application into microservices and use Amazon DynamoDB for data storage.
Why wrong: High refactoring effort; not directly reducing licensing costs.
C
Replatform the database to Amazon RDS for PostgreSQL and refactor the application to use it.
Reduces licensing costs and improves scalability.
D
Rehost the application on Amazon EC2 and use Amazon RDS for Oracle with license-included.
Why wrong: License-included is more expensive.
E
Rehost the application on Amazon EC2 and use Amazon RDS for Oracle with BYOL.
Why wrong: Still uses Oracle licensing; does not reduce costs.

Question 7mediummatching

Match each AWS cost management tool to its use.

Drag a concept onto its matching description — or click a concept then click the description.

Concepts

Matches

Visualize and explore cost and usage data

Set custom cost and usage budgets with alerts

Recommendations for cost optimization, performance, security

Flexible pricing model for compute savings

Recommend optimal compute resources based on usage

Question 8mediummatching

Match each AWS disaster recovery strategy to its description.

Drag a concept onto its matching description — or click a concept then click the description.

Concepts

Matches

Lowest cost, RPO in hours, RTO in hours to days

Core services run in standby, ready to scale

Scaled-down production environment, ready to scale up

Both sites serve traffic, failover is immediate

One site active, other on standby

Question 9mediummultiple choice

A company's AWS environment includes multiple VPCs across several accounts that are connected via a transit gateway. The network team wants to monitor all network traffic between VPCs for security analysis. Which solution is the most scalable and cost-effective?

Trap 1: Use AWS Traffic Mirroring to mirror all traffic to a central…

High cost and complexity.

Trap 2: Enable VPC Flow Logs for each VPC and stream them to Amazon…

Logs are not centralized.

Trap 3: Place a network load balancer in each VPC and capture traffic using…

Not designed for traffic monitoring.

A
Use AWS Traffic Mirroring to mirror all traffic to a central inspection appliance.
Why wrong: High cost and complexity.
B
Enable VPC Flow Logs and publish them to a central S3 bucket, then use Amazon Athena to query the logs.
Scalable and cost-effective.
C
Enable VPC Flow Logs for each VPC and stream them to Amazon CloudWatch Logs in each account.
Why wrong: Logs are not centralized.
D
Place a network load balancer in each VPC and capture traffic using a packet sniffer.
Why wrong: Not designed for traffic monitoring.

Question 10hardmultiple choice

A company has a central logging account that receives VPC Flow Logs, CloudTrail logs, and AWS Config logs from all accounts in the organization. The logs are stored in S3 buckets. The security team wants to analyze these logs using Amazon Athena. What is the MOST cost-effective way to ensure that the Athena queries only scan the necessary data?

Trap 1: Use S3 object-level compression (e.g., gzip) to reduce data volume.

Compression reduces storage but does not reduce the amount of data scanned if partitions are not used.

Trap 2: Create AWS Glue partition indexes on the table.

Partition indexes are for faster partition retrieval, not for limiting scan size.

Trap 3: Create separate Athena tables for each account and region.

This increases maintenance and query complexity.

A
Partition the data by account ID, region, and date in the S3 bucket, and use partitions in Athena.
Partition pruning ensures Athena scans only relevant partitions.
B
Use S3 object-level compression (e.g., gzip) to reduce data volume.
Why wrong: Compression reduces storage but does not reduce the amount of data scanned if partitions are not used.
C
Create AWS Glue partition indexes on the table.
Why wrong: Partition indexes are for faster partition retrieval, not for limiting scan size.
D
Create separate Athena tables for each account and region.
Why wrong: This increases maintenance and query complexity.

Question 11easymultiple choice

A company uses AWS Organizations with consolidated billing. The finance team wants to track costs by project, where each project may use resources across multiple accounts. What is the MOST scalable way to allocate costs?

Trap 1: Create separate AWS accounts for each project and use consolidated…

This is not scalable for many projects and does not allow cross-account resource sharing.

Trap 2: Use AWS Budgets with cost allocation tags to track project costs.

Budgets alert but do not provide aggregated cost allocation across accounts.

Trap 3: Apply cost allocation tags to resources in each account and use AWS…

Tags must be applied consistently across accounts, but the activation and reporting can be centralized.

A
Create separate AWS accounts for each project and use consolidated billing to view costs.
Why wrong: This is not scalable for many projects and does not allow cross-account resource sharing.
B
Use AWS Budgets with cost allocation tags to track project costs.
Why wrong: Budgets alert but do not provide aggregated cost allocation across accounts.
C
Apply cost allocation tags to resources in each account and use AWS Cost Explorer to filter by tag.
Why wrong: Tags must be applied consistently across accounts, but the activation and reporting can be centralized.
D
Use AWS Cost Categories in AWS Cost Explorer to group costs by shared tags that are activated centrally.
Cost Categories allow grouping by tags, accounts, or services and are managed centrally in the management account.

Question 12hardmultiple choice

A company is deploying a multi-tier web application on AWS. The application must be highly available across three Availability Zones. The web tier runs on EC2 instances behind an Application Load Balancer (ALB). The application tier runs on EC2 instances behind a Network Load Balancer (NLB). The database tier uses a Multi-AZ RDS instance. To reduce cross-AZ data transfer costs, which design should be implemented?

Trap 1: Use AWS Global Accelerator to reduce data transfer costs

Global Accelerator improves performance but does not reduce internal cross-AZ costs.

Trap 2: Place all web tier instances in one AZ and all application tier…

This increases cross-AZ traffic and costs.

Trap 3: Use a single AZ for all tiers to avoid cross-AZ traffic

Single AZ reduces high availability.

A
Use AWS Global Accelerator to reduce data transfer costs
Why wrong: Global Accelerator improves performance but does not reduce internal cross-AZ costs.
B
Place all web tier instances in one AZ and all application tier instances in another AZ
Why wrong: This increases cross-AZ traffic and costs.
C
Use a single AZ for all tiers to avoid cross-AZ traffic
Why wrong: Single AZ reduces high availability.
D
Place web and application tier instances in the same subnets across all three AZs
ALB and NLB can route to targets in the same AZ, reducing cross-AZ traffic.

Question 13mediummultiple choice

A company has multiple AWS accounts and wants to centralize operational data such as cost reports, security findings, and resource inventory. The solution should provide a single pane of glass for the operations team. Which AWS service should be used?

Trap 1: Use Amazon QuickSight to connect to each account's AWS Cost and…

QuickSight can visualize but does not aggregate automatically.

Trap 2: Use AWS Control Tower to set up a landing zone and use the built-in…

Control Tower dashboard is limited to governance, not operational data.

Trap 3: Use AWS Config aggregator to collect resource configurations from…

Config aggregator only collects config data, not cost or security.

A
Use Amazon QuickSight to connect to each account's AWS Cost and Usage Report and AWS Config data.
Why wrong: QuickSight can visualize but does not aggregate automatically.
B
Use AWS Control Tower to set up a landing zone and use the built-in dashboard for operational data.
Why wrong: Control Tower dashboard is limited to governance, not operational data.
C
Use AWS Config aggregator to collect resource configurations from all accounts.
Why wrong: Config aggregator only collects config data, not cost or security.
D
Use AWS Organizations to create a central management account and use AWS Systems Manager Explorer to aggregate operational data.
Systems Manager Explorer provides a single dashboard for operational data.

Question 14mediummultiple choice

A company is using AWS Organizations with multiple accounts. The central IT team wants to enforce that all EC2 instances are launched with specific tags (e.g., CostCenter and Environment). The solution should prevent any untagged instances from being created. Which approach should be taken?

Trap 1: Use AWS Service Catalog to provision EC2 instances only from…

This does not prevent users from launching instances outside Service Catalog.

Trap 2: Use an AWS Config rule to detect untagged instances and trigger a…

This is reactive, not preventive.

Trap 3: Create an IAM policy for each account that requires tags on…

IAM policies cannot enforce tags at launch time; they can only check condition keys.

A
Use AWS Service Catalog to provision EC2 instances only from pre-configured products that include required tags.
Why wrong: This does not prevent users from launching instances outside Service Catalog.
B
Use an AWS Config rule to detect untagged instances and trigger a Lambda function to terminate them.
Why wrong: This is reactive, not preventive.
C
Create an SCP that denies the ec2:RunInstances action if the required tags are not specified in the request.
SCPs can enforce tagging requirements at the organization level.
D
Create an IAM policy for each account that requires tags on instance creation.
Why wrong: IAM policies cannot enforce tags at launch time; they can only check condition keys.

Question 15easymultiple choice

A solutions architect is designing a new serverless application using AWS Lambda for business logic, Amazon API Gateway for RESTful APIs, and Amazon DynamoDB for data storage. The application will experience unpredictable traffic spikes. What is the MOST cost-effective way to handle concurrency and scaling?

Trap 1: Use Lambda provisioned concurrency to pre-warm instances.

Provisioned concurrency adds cost and is not needed for unpredictable spikes; it's for latency-sensitive applications.

Trap 2: Configure DynamoDB auto scaling to handle traffic spikes.

DynamoDB auto scaling manages read/write capacity, not Lambda concurrency.

Trap 3: Set a usage plan in API Gateway with a throttling limit.

API Gateway throttling limits API requests but does not directly control Lambda concurrency.

A
Use Lambda provisioned concurrency to pre-warm instances.
Why wrong: Provisioned concurrency adds cost and is not needed for unpredictable spikes; it's for latency-sensitive applications.
B
Use Lambda reserved concurrency to set a limit on concurrent executions.
Reserved concurrency controls the maximum number of concurrent Lambda invocations, preventing excessive scaling and cost.
C
Configure DynamoDB auto scaling to handle traffic spikes.
Why wrong: DynamoDB auto scaling manages read/write capacity, not Lambda concurrency.
D
Set a usage plan in API Gateway with a throttling limit.
Why wrong: API Gateway throttling limits API requests but does not directly control Lambda concurrency.

Question 16easymultiple choice

A company needs to design a new solution for storing and retrieving user-uploaded images. The images are accessed frequently for the first 30 days and then rarely accessed after that. The company wants to minimize storage costs while maintaining low-latency access for frequently accessed images. Which storage strategy should be used?

Trap 1: Store images in Amazon EBS volumes attached to a web server.

EBS is block storage and not designed for object storage at scale; also cost-inefficient.

Trap 2: Store all images in Amazon S3 Glacier Flexible Retrieval.

Glacier has retrieval times that don't meet low-latency requirements for first 30 days.

Trap 3: Store all images in Amazon S3 Standard.

S3 Standard is costly for long-term storage of rarely accessed data.

A
Store images in Amazon EBS volumes attached to a web server.
Why wrong: EBS is block storage and not designed for object storage at scale; also cost-inefficient.
B
Store images in Amazon S3 Standard and use lifecycle policies to transition to S3 Standard-IA after 30 days.
Lifecycle policies automate cost optimization while keeping low-latency access during frequent access period.
C
Store all images in Amazon S3 Glacier Flexible Retrieval.
Why wrong: Glacier has retrieval times that don't meet low-latency requirements for first 30 days.
D
Store all images in Amazon S3 Standard.
Why wrong: S3 Standard is costly for long-term storage of rarely accessed data.

Read the full NAT/PAT explanation →

Question 17mediummulti select

Which TWO strategies can be used to reduce the cost of Amazon DynamoDB tables for a new application with unpredictable traffic patterns? (Choose two.)

Trap 1: Use DynamoDB Streams to reduce write capacity.

Streams are used for change data capture, not for reducing capacity.

Trap 2: Use DynamoDB global tables for multi-region replication.

Global tables increase cost due to replication.

Trap 3: Use DynamoDB Accelerator (DAX) to reduce read capacity.

DAX is a caching layer that reduces read latency but does not directly reduce cost of the table; it adds cost.

A
Use DynamoDB auto scaling with provisioned capacity.
Auto scaling adjusts capacity based on usage, preventing over-provisioning.
B
Use DynamoDB Streams to reduce write capacity.
Why wrong: Streams are used for change data capture, not for reducing capacity.
C
Use DynamoDB global tables for multi-region replication.
Why wrong: Global tables increase cost due to replication.
D
Use DynamoDB Accelerator (DAX) to reduce read capacity.
Why wrong: DAX is a caching layer that reduces read latency but does not directly reduce cost of the table; it adds cost.
E
Use DynamoDB on-demand capacity mode.
On-demand mode charges per request, suitable for unpredictable traffic.

Question 18easymultiple choice

A company wants to design a cost-effective solution to store infrequently accessed log files for 7 years. The logs are generated daily and must be available for retrieval within 24 hours. Which Amazon S3 storage class should be used?

Trap 1: S3 One Zone-Infrequent Access

Not designed for 7-year retention.

Trap 2: S3 Intelligent-Tiering

Not cost-effective for known infrequent access pattern.

Trap 3: S3 Standard

Too expensive for long-term archival.

A
S3 One Zone-Infrequent Access
Why wrong: Not designed for 7-year retention.
B
S3 Intelligent-Tiering
Why wrong: Not cost-effective for known infrequent access pattern.
C
S3 Glacier Deep Archive
Lowest cost with 12-24 hour retrieval.
D
S3 Standard
Why wrong: Too expensive for long-term archival.

Question 19easymulti select

A company is designing a cost-effective architecture for a batch processing job that runs nightly. The job can tolerate interruptions and requires significant compute power for a few hours. The company wants to minimize costs. Which TWO strategies should the company use?

Trap 1: Purchase Reserved Instances (RI) for a 1-year term.

RIs require upfront commitment and are not ideal for variable workloads.

Trap 2: Use On-Demand Instances to ensure availability.

On-Demand is more expensive.

Trap 3: Use Dedicated Hosts for compliance.

Dedicated Hosts are costly and unnecessary.

A
Use Spot Instances for compute.
Spot Instances are cost-effective and suitable for fault-tolerant batch jobs.
B
Purchase Reserved Instances (RI) for a 1-year term.
Why wrong: RIs require upfront commitment and are not ideal for variable workloads.
C
Configure Auto Scaling to scale out during the job and scale in after.
Auto Scaling ensures you only pay for what you use.
D
Use On-Demand Instances to ensure availability.
Why wrong: On-Demand is more expensive.
E
Use Dedicated Hosts for compliance.
Why wrong: Dedicated Hosts are costly and unnecessary.

Question 20easymultiple choice

A company wants to serve static content (images and videos) to users worldwide with low latency. The content is stored in an Amazon S3 bucket. What is the most cost-effective solution?

Trap 1: Use AWS Global Accelerator with endpoints pointing to the S3 bucket.

Global Accelerator improves network routing but does not cache content; CloudFront is the appropriate CDN.

Trap 2: Deploy EC2 instances in multiple Regions and use a load balancer.

This adds complexity and cost; CloudFront is simpler and cost-effective for static content.

Trap 3: Host the content directly from the S3 bucket and use S3 Transfer…

Transfer Acceleration speeds up uploads, not downloads; it does not cache content globally.

A
Use AWS Global Accelerator with endpoints pointing to the S3 bucket.
Why wrong: Global Accelerator improves network routing but does not cache content; CloudFront is the appropriate CDN.
B
Deploy EC2 instances in multiple Regions and use a load balancer.
Why wrong: This adds complexity and cost; CloudFront is simpler and cost-effective for static content.
C
Use Amazon CloudFront with the S3 bucket as the origin.
CloudFront caches content at edge locations, reducing latency and data transfer costs.
D
Host the content directly from the S3 bucket and use S3 Transfer Acceleration.
Why wrong: Transfer Acceleration speeds up uploads, not downloads; it does not cache content globally.