AWS Certified Data Engineer Associate DEA-C01 Practice Question

An analytics team needs to collect clickstream events from thousands of retail websites. The pipeline must sustain 20 MB/s and absorb short-lived spikes up to 200 MB/s without pre-provisioning capacity. Downstream consumers reading the data need sub-100 ms latency and the ability to replay any event that arrived during the previous 7 days. The team wants a fully managed ingestion service that requires no cluster administration. Which solution meets these requirements?

Use Amazon Kinesis Data Streams in on-demand mode with the retention period set to 7 days.
Enable Amazon DynamoDB Streams on a table that stores the events and configure extended retention.
Deploy an Amazon MSK cluster with auto-scaling partitions and a 7-day retention policy.
Use Amazon Kinesis Data Firehose to deliver the events directly to Amazon S3 with buffering set to 60 seconds.

Report Issue

Answer Description

Amazon Kinesis Data Streams in on-demand mode removes capacity planning. A new on-demand stream can automatically scale up to 200 MB/s write and 400 MB/s read throughput per stream by default and higher upon request. The service supports a configurable retention period of up to 365 days, so a 7-day replay window is easily satisfied. With enhanced fan-out consumers, records are typically delivered to readers within about 70 milliseconds-well under the 100-ms requirement. If traffic grows by more than 2× the previous peak within 15 minutes, some writes may be throttled, but standard retry logic absorbs those temporary errors without capacity pre-provisioning.

Amazon Data Firehose (formerly Kinesis Data Firehose) buffers data for seconds before delivery and offers no built-in replay capability, so it fails the latency and replay requirements. Amazon MSK provisioned clusters still require choosing broker instance types, counts, and patching, which contradicts the "no administration" goal. DynamoDB Streams is intended for table change capture, retains data only 24 hours, and cannot sustain the required throughput.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.