DEA-C02 Question #38: Real Exam Question with Answer & Explanation

Question

A Data Engineer is working on a continuous data pipeline which receives data from Amazon Kinesis Firehose and loads the data into a staging table which will later be used in the data transformation process. The average file size is 300-500 MB. The Engineer needs to ensure that Snowpipe is performant while minimizing costs. How can this be achieved?

Options

• AIncrease the size of the virtual warehouse used by Snowpipe.
• BSplit the files before loading them and set the SIZE_LIMIT option to 250 MB.
• CChange the file compression size and increase the frequency of the Snowpipe loads.
• DDecrease the buffer size to trigger delivery of files sized between 100 to 250 MB in Kinesis

Explanation

Adjusting Kinesis Firehose's buffer size is the correct approach because Snowpipe is optimized for files in the 100–250 MB range - files of 300–500 MB are too large, causing slower ingestion and higher compute costs. Since Kinesis Firehose is the upstream source, configuring its buffer settings (size and interval) is the most direct way to control the file size that lands in S3 before Snowpipe picks it up.

Why the distractors are wrong:

• A is wrong because Snowpipe is serverless - it does not use a virtual warehouse that you can resize; it manages its own compute automatically.
• B is wrong because SIZE_LIMIT in a Snowpipe COPY command caps the amount of data loaded per run, not the physical file size - it doesn't split files and adds unnecessary complexity.
• C is wrong because changing compression format and increasing load frequency doesn't fix the oversized-file problem and could actually increase costs by triggering more Snowpipe invocations.

Memory tip: Think of Snowpipe as having a "Goldilocks zone" of 100–250 MB per file. When files arrive too big, tune the source buffer (Kinesis), not the destination - Kinesis is the valve that controls the size of what gets delivered.

No community discussion yet for this question.