Practice Test 1 | Google Cloud Certified Professional Data Engineer | Dumps | Mock Test
A social media platform stores various details of their platform users such as session login time, URLs visited, activities on platform and other logs. With GDPR (General Data Protection Regulation) compliance to be officially implemented, the platform now allows users to download their activity logs from their profile settings which they can click a button to call an API to generate a full report.
Recently, users are complaining timeouts after 60 seconds of requesting to download their activity logs at peak hours when the platform has the most traffic. They have to try for several minutes or even hours for the API to return their report available for download.
How can you solve this issue?
A. Increase timeout for API at peak times to 120 seconds. If it keeps failing, try increasing the timeout until the issue is resolved.
B. Build a Dataflow pipeline to generate daily reports of users’ activity logs. Users can download those daily reports whenever they want to.
C. Migrate data source to Cloud Spanner for horizontal scaling to avoid query timeouts.
D. Use Pub/sub to pull the requests for activity logs from users. Send a link to users by their email addresses with a temporary download link for them to access their report.
Answer: D.
Cloud Pub/Sub is a service to ingest event streams at any scale. It’s scalable and reliable for stream analytics and event-driven computing systems.
Pub/sub is a good product to de-couple a system’s components so they communicate with each other asymmetrically. From the scenario shown here, instead of directly calling the API to export required report which puts great loads on the API and hence the timeouts faced by users. Instead, the platform can “publish” messages to a “topic” related to exporting activity log reports sending the required parameters such as user ID and custom settings such as date range and what data to export. The API can be switched to be a “subscriber” which receives the messages sent and processes each message asymmetrically to generate the report, then sends the download link to the user’s mailbox when ready.
Hence, answer D is correct.
Answer A is incorrect: Increasing timeout isn’t a scalable solution and it may keep occurring eventually when more and more users join the platform.
Answer B is incorect: While this would solve the timeout issues, generating daily reports for users can be costly as more users join, knowing that requesting activity log reports are a non-frequent action and this costs both compute and storage resources. This solution also doesn’t provide flexibility with what parameters the report is generated on such as date range and other custom metrics.
Answer C is incorrect: This solution has several issues. First, we’re assuming the data source is a relational database, which can be unlikely since NoSQL databases better perform for massive log input which uses the user ID as a key to reach the data. Second, Cloud Spanner isn’t a cheap solution for a service not frequently used.
Source(s):
Google Pub/Sub: https://cloud.google.com/pubsub/docs/overview
Comments are closed, but trackbacks and pingbacks are open.