Design High-Performing Architectures24% · 5 topics

Q1. A gp2 EBS volume has enough capacity but not enough IOPS. Which is the most appropriate way to increase performance while keeping costs down?

1. Change the volume to gp3 and set the required IOPS separately from capacity
2. Keep using gp2 with more capacity than needed just to get IOPS
3. Change the volume to sc1 to lower cost
4. Stop using EBS and put all data on instance store

Q2. An app on EC2 is slow. CloudWatch shows CPU utilization is low, but EBS volume reads and writes are pinned at the limit. What should you do next?

1. Increase the Auto Scaling maximum
2. Delete the CloudWatch alarms
3. Review the EBS volume type and IOPS/throughput settings and change them as needed
4. Change to a larger instance type to add CPU and see what happens

Q1. To speed up batch processing, you want to split it into many small jobs and run them in parallel. Which approach keeps infrastructure management to a minimum?

1. Keep an EMR cluster always running and submit jobs
2. Run jobs in parallel with AWS Batch or a Step Functions distributed map
3. Create an EC2 Auto Scaling group and distribute jobs with scripts
4. Write all jobs as parallel threads inside one Lambda function

Q2. A Lambda function takes a long time to run and has spare memory. Which is the most appropriate way to make it faster?

1. Recreate the function in another Region
2. Shorten the timeout to cut processing early and leave the rest for next time
3. Limit concurrency to 1
4. Increase the function's memory (CPU increases proportionally)

Q1. Read load on RDS for PostgreSQL is high and responses are slow. Writes are few. Which is the most appropriate way to improve performance with minimal app changes?

1. Add read replicas and direct read-only queries to them
2. Stop the DB instance once and reduce its allocated storage
3. Send read queries to the Multi-AZ standby
4. Shorten the database backup retention period

Q2. A relational DB has a large gap between quiet periods and sudden peaks, and you want to automate capacity adjustment. Which is most appropriate?

1. Use Aurora replica Auto Scaling to change the replica count
2. Use Aurora Serverless v2 to scale capacity automatically with load
3. Enable RDS storage autoscaling
4. Add RDS Proxy to pool connections and reduce load

Q1. Requests to one S3 prefix are extremely numerous and approaching the performance limit. Which design improvement is most appropriate?

1. Move the bucket to S3 Express One Zone and keep the same prefix
2. Put CloudFront in front and cache write requests too
3. Enable S3 Transfer Acceleration
4. Spread object keys across multiple prefixes and send requests in parallel

Q2. You want to transfer large amounts of data daily between the head office and the Tokyo Region with stable bandwidth, unaffected by internet congestion. Which is most appropriate?

1. S3 Transfer Acceleration
2. Amazon CloudFront
3. AWS Direct Connect
4. AWS Site-to-Site VPN (over the internet)

Q1. Kinesis Data Streams processing cannot keep up, and ingestion is being throttled. Which measure is most appropriate?

1. Shorten the data retention period
2. Reduce the number of producers and stop sending
3. Delete the stream and have all producers write directly to S3
4. Increase the number of shards (or use on-demand mode)

Q2. You are comparing analytics building blocks. Which statement about Amazon Athena is incorrect?

1. Data cannot be queried without loading it into dedicated storage
2. It is billed by the amount of data scanned by queries
3. It can query data in S3 with standard SQL
4. No server management is required