MLS-C01 Question #139: Real Exam Question with Answer & Explanation

Question

A company ingests machine learning (ML) data from web advertising clicks into an Amazon S3 data lake. Click data is added to an Amazon Kinesis data stream by using the Kinesis Producer Library (KPL). The data is loaded into the S3 data lake from the data stream by using an Amazon Kinesis Data Firehose delivery stream. As the data volume increases, an ML specialist notices that the rate of data ingested into Amazon S3 is relatively constant. There also is an increasing backlog of data for Kinesis Data Streams and Kinesis Data Firehose to ingest. Which next step is MOST likely to improve the data ingestion rate into Amazon S3?

Options

• AIncrease the number of S3 prefixes for the delivery stream to write to.
• BDecrease the retention period for the data stream.
• CIncrease the number of shards for the data stream.
• DAdd more consumers using the Kinesis Client Library (KCL).

Explanation

{"question_number": 5, "correct_answer": "A", "explanation": "The symptom - a constant (capped) S3 ingestion rate despite a growing backlog in both Kinesis Data Streams and Kinesis Data Firehose - points to a bottleneck at the Amazon S3 write layer, not at the Kinesis stream layer. Amazon S3 enforces per-prefix request rate limits (3,500 PUT/COPY/POST/DELETE requests per second per prefix). When Kinesis Data Firehose is writing to a single prefix, it hits this ceiling. By increasing the number of S3 prefixes in the Firehose delivery configuration, Firehose can distribute writes across multiple prefixes in parallel, multiplying the effective write throughput into S3. Option B (decreasing retention period) reduces data availability but does not increase write speed. Option C (increasing shards) improves Kinesis Data Streams throughput, but if S3 is already saturated, more shards won't help the downstream bottleneck. Option D (adding KCL consumers) adds read capacity from the stream, not write capacity into S3.", "generated_by": "claude-sonnet", "llm_judge_score": 5}

No community discussion yet for this question.