GCP Professional Data Engineer Practice Question

A retailer runs a Dataflow streaming pipeline that consumes customer events from Pub/Sub and enriches them with product data stored in BigQuery. The analytics team now needs to add near-real-time inventory information coming from an on-premises Oracle 12c database. The solution must capture change data without installing agents on the source database server to avoid performance impact, deliver the change stream to the existing Dataflow job with seconds-level latency, and minimize custom code and ongoing operational overhead. What should you do?

Use Database Migration Service in continuous replication mode to export Oracle change logs to Cloud Storage and schedule a nightly Dataflow batch job that loads incremental files into BigQuery.
Set up BigQuery connection federation to the on-premises Oracle instance and reference the external tables directly from the existing Dataflow SQL transform.
Deploy an open-source Debezium cluster on Google Kubernetes Engine to read Oracle changes and write them to a Kafka cluster; create a new Kafka I/O transform in Dataflow to consume the data.
Configure Datastream to capture Oracle redo logs, route the stream through the Datastream-to-Pub/Sub template, and add the new Pub/Sub subscription to the existing Dataflow pipeline for real-time enrichment.

Report Issue

Answer Description

Datastream is Google Cloud's serverless change data capture (CDC) service that can continuously read Oracle redo logs without agents, imposing minimal load on the source. It writes change events to Cloud Storage, and the Datastream-to-Pub/Sub Dataflow template streams every event to a Pub/Sub topic. By adding this topic as an additional input to the existing Dataflow pipeline, you meet the latency, agentless, and low-operations requirements. Database Migration Service does not stream Oracle changes into Pub/Sub, Debezium on GKE introduces significant operational burden, and BigQuery federation does not provide streaming CDC, so these alternatives do not satisfy the constraints.

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.