You are a developer at a company that operates an ecommerce website. The website stores the customer order data in a Cloud SQL for PostgreSQL database. Data scientists on the marketing team access…

Question

Options

• AConfigure Cloud Scheduler to run an hourly Cloud Run function that exports the data from the
• BSet up a Bigtable table for the data science team. Configure the application to perform dual writes
• CSet up a BigQuery dataset for the data science team. Configure Datastream to replicate the
• DCreate a clone of the PostgreSQL database instance for the data science team. Schedule a job to

• A
  7% (3)
• B
  4% (2)
• C
  73% (33)
• D
  16% (7)

Explanation

Option C - using Datastream to replicate the Cloud SQL PostgreSQL data into a BigQuery dataset - is the optimal solution because: (1) Datastream performs low-latency, continuous change data capture (CDC) replication, so data scientists always have near-real-time data without polling the production database. (2) BigQuery is purpose-built for large-scale analytical queries, so report workloads run efficiently there without competing with transactional OLTP workloads on Cloud SQL. Option A (hourly export via Cloud Run) introduces up to a 1-hour data lag and adds unnecessary complexity. Option B (Bigtable dual-writes) adds application code complexity and Bigtable is not optimized for SQL-style analytical reporting. Option D (cloning the database) still requires the clone to be kept in sync and does not scale analytically as well as BigQuery.

No community discussion yet for this question.