GCP Professional Data Engineer Practice Question

Your company ingests 4 TB of raw click-stream files into a Cloud Storage bucket every day. Data scientists only analyze the most recent 14 days of data, while the rest must be retained for 13 months to satisfy quarterly compliance audits. Historical data might be accessed at most a few times per year and can tolerate several hours of retrieval latency. To minimize storage costs, what should you do?

Load all files into BigQuery, relying on automatic long-term storage pricing after 90 days for cost reduction.
Keep all objects in the existing bucket using the Standard storage class to avoid retrieval fees.
Move objects older than 14 days to a Filestore instance mounted on persistent SSDs, starting and stopping the instance when needed.
Use an object lifecycle rule to retain the first 14 days in Standard storage and transition older objects to the Archive storage class for the remainder of the retention period.

Report Issue

Answer Description

Archive storage is the lowest-cost Cloud Storage class and is designed for data that is retained for long periods and accessed infrequently; objects can be retrieved within hours, which matches the tolerance described. Keeping the newest 14 days in the Standard class avoids per-GB retrieval charges and provides low-latency access for the analysts. A lifecycle rule that automatically transitions objects from Standard to Archive after 14 days eliminates manual effort and prevents paying Standard rates for terabytes that are rarely read. Leaving everything in Standard would keep costs unnecessarily high. Loading the data into BigQuery still incurs higher monthly storage fees than Archive and adds query costs. Moving data to Filestore places it on persistent disks priced significantly above Archive storage and requires an always-running instance, further increasing expense.

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.