Onboarding a data source the pipeline has not seen before

Integrating with new data sources on Google Cloud requires selecting the right ingestion pattern (batch, streaming, or change data capture) based on source type, volume, and latency needs, then choosing an appropriate connector or service to land data reliably. Google Cloud provides managed services like Datastream, Pub/Sub, Storage Transfer Service, BigQuery Data Transfer Service, and Dataflow templates to reduce custom integration code, while Cloud Data Fusion and Apiary/Apigee-style API integrations extend connectivity to SaaS and on-premises systems. Understanding source constraints (schema, network access, authentication) and target requirements (schema evolution, deduplication, exactly-once semantics) is key to designing a robust pipeline.

Must-know

  • Datastream provides serverless change data capture (CDC) and replication from Oracle, MySQL, PostgreSQL, and AlloyDB into BigQuery, Cloud Storage, or Cloud SQL with minimal source impact.
  • Pub/Sub is the standard entry point for streaming and event-driven integrations, decoupling producers and consumers and supporting push/pull delivery with at-least-once guarantees.
  • Storage Transfer Service handles large-scale batch data movement from on-premises, other clouds (S3, Azure Blob), or HTTP/HTTPS sources into Cloud Storage, including scheduled and incremental transfers.
  • BigQuery Data Transfer Service automates recurring loads from SaaS applications (Google Ads, YouTube, etc.) and other data warehouses (Teradata, Redshift, S3) directly into BigQuery.
  • Cloud Data Fusion offers a visual, no/low-code interface with hundreds of pre-built connectors (via CDAP plugins) for integrating diverse enterprise and cloud data sources into pipelines.
  • For custom or unsupported sources, Dataflow (with Apache Beam I/O connectors) or a Pub/Sub-based ingestion pattern combined with Cloud Functions/Cloud Run is the typical fallback for building bespoke connectors, and network connectivity (VPC peering, Private Service Connect, or Cloud VPN/Interconnect) must be considered for on-premises or hybrid sources.
Drill this objectiveTake a weighted mockTry free practice questions

Coverage checked against the published exam guide on Aug 11, 2026.

These are independent practice questions, written against this certification's published exam guide. They are not the certification vendor's own questions, and not the real exam.