01. Every night, a job on EC2 instances in Region B reads about 5 TB of new log files from an S3 bucket in Region A, aggregates them, and writes a summary of a few gigabytes back to the same bucket. Data transfer between Regions is now the largest item on the bill. The log files must stay in Region A.
Which change will reduce the data transfer cost the MOST?
a) Add an S3 gateway endpoint to the VPC in Region B.
b) Enable S3 Transfer Acceleration on the bucket.
c) Replicate the bucket to Region B with S3 Cross-Region Replication.
d) Run the job on EC2 instances in Region A.
02. A company stores product photos in the S3 Standard storage class. A recommendation engine reads most photos several times each week for as long as the photos are kept. To cut storage cost, a team proposes a lifecycle rule that moves every photo to S3 Standard-IA 30 days after upload.
What should a solutions architect recommend?
a) Apply the rule but target S3 One Zone-IA instead
b) Apply the rule with the transition delayed to 90 days
c) Keep the photos in S3 Standard with no transition
d) Apply the rule but target S3 Glacier Instant Retrieval
03. A company has 40 VPCs in one AWS Region, all attached to a transit gateway. Each VPC has its own interface VPC endpoints for the same six AWS services, provisioned in two Availability Zones. Traffic to these endpoints is light, and the endpoint charges now far exceed what the traffic would justify.
Workloads must keep reaching the services privately, without traversing the internet, and must keep Availability Zone resilience.
Which solution will reduce cost the MOST while meeting these requirements?
a) Replace the interface endpoints in each VPC with gateway endpoints for the same six services.
b) Keep the endpoints in each VPC but provision them in only one Availability Zone.
c) Remove the endpoints and send the service traffic through a NAT gateway in each VPC.
d) Host one set of the interface endpoints in a shared services VPC, and resolve them from the other VPCs through Route 53 private hosted zones over the transit gateway.
04. A DynamoDB table in provisioned capacity mode has plenty of unused write capacity, yet writes to the table are being throttled. The table has a global secondary index whose key attributes are present on every item, and the index was given much less provisioned write capacity than the table.
What should the solutions architect do to stop the throttling?
a) Increase the provisioned write capacity of the table.
b) Raise the global secondary index’s write capacity.
c) Add a DynamoDB Accelerator (DAX) cluster in front of the table.
d) Increase the provisioned read capacity of the global secondary index.
05. An analytics web application runs on Amazon EC2 instances in an Auto Scaling group. Each new instance runs a startup script that downloads and indexes a large dataset, and an instance is not ready to serve requests until about 15 minutes after launch.
Traffic spikes arrive at unpredictable times, and users see slow responses until the new instances are ready.
Which solution will MOST reduce the time it takes for new capacity to start serving requests?
a) Add a warm pool of pre-initialized instances to the group
b) Turn on predictive scaling based on historical load
c) Set a longer default instance warmup for the Auto Scaling group
d) Add a scheduled action that raises the desired capacity every morning
06. A company runs a mix of Amazon EC2 instances, AWS Fargate tasks and AWS Lambda functions. Its usage has been stable for six months, and the finance team has decided to buy a Compute Savings Plan.
A solutions architect must recommend the hourly commitment amount, based on the company's actual usage, with the LEAST effort.
What should the solutions architect use?
a) AWS Compute Optimizer rightsizing recommendations for the EC2 instances
b) Savings Plans recommendations in the Billing and Cost Management console
c) The AWS Cost and Usage Report, queried with Amazon Athena
d) An AWS Budgets cost budget with a forecasted-spend alert
07. A company backs up on-premises servers with backup software that writes to a physical tape library. Tapes are shipped offsite and kept for 10 years but are almost never restored. The company wants to stop handling physical tapes, keep its existing backup software and jobs, and pay the lowest storage cost for the archived tapes.
Which solution meets these requirements?
a) Deploy S3 File Gateway and add a lifecycle rule to S3 Glacier Flexible Retrieval.
b) Deploy Tape Gateway and archive ejected tapes to S3 Glacier Deep Archive.
c) Deploy Volume Gateway in stored mode and take EBS snapshots of the volumes.
d) Use AWS DataSync to copy the backup files to S3 Glacier Instant Retrieval.
08. A company will run a database server on one Amazon EC2 instance of a fixed instance type in one Region, 24 hours a day, for at least the next three years. The instance must not be interrupted, and the company does not expect to change the instance family. Paying the full amount in advance is acceptable.
Which purchase option gives the LOWEST cost for this instance?
a) A Convertible Reserved Instance with a 3-year term and All Upfront payment.
b) A Standard Reserved Instance with a 1-year term, renewed each year.
c) A Standard Reserved Instance with a 3-year term and All Upfront payment.
d) Spot Instances with a maximum price set at the On-Demand rate.
09. An Amazon Aurora MySQL-Compatible DB cluster has a writer instance and three Aurora Replicas. One replica uses the same instance class as the writer; the other two are smaller and serve reporting queries. If the writer fails, the larger replica must become the new writer so that performance does not degrade.
What should a solutions architect do?
a) Assign the larger replica a higher-priority promotion tier than the smaller replicas
b) Create an RDS Proxy and set the larger replica as its default target
c) Configure the cluster reader endpoint to include only the larger replica
d) Add the larger replica to a custom endpoint and point the application's write traffic at it
10. An order service sends messages to an Amazon SQS standard queue that uses its default settings. Worker instances poll the queue, process each message, and then delete it. Processing one message takes about four minutes.
The operations team finds that many orders are processed twice, and the workers log no errors.
What should a solutions architect do to stop the repeated processing?
a) Turn on long polling by setting the receive message wait time on the queue to its maximum.
b) Add a delivery delay of several minutes to every message sent to the queue.
c) Set the visibility timeout longer than a worker needs to process and delete a message.
d) Increase the queue's message retention period to its maximum value.