Simulation
Change the workload source and processing requirement to see which ingestion path fits.
Change the inputs and watch how the outcome changes.

Change the workload source and processing requirement to see which ingestion path fits.
With an event stream that needs complex transforms, Pub/Sub ingests and buffers asynchronous bursts from many producers, and Dataflow autoscales worker VMs to run Apache Beam filtering, aggregation, joins, or ML inference before writing to BigQuery. This is the exam's choice when the stream needs real processing before loading.
Starting scenarioEvery input at its default
Because the requirement changed to a direct load with at most a simple UDF, the stream bypasses Dataflow entirely: a Pub/Sub BigQuery subscription streams events straight into BigQuery tables. This minimizes operational complexity and latency when no complex transforms, joins, or enrichment are needed.
What changed
The workload source changed from an event stream to operational database changes, and the load is direct, so Datastream is the fit: its serverless change data capture captures ongoing row-level changes and replicates them straight into BigQuery with minimal latency, with no intermediate processing tier.
What changed
Database changes still come from Datastream CDC, but the complex-processing requirement routes them through Cloud Storage first: Datastream writes the change event stream as Avro or JSON files, then Dataflow templates load, enrich, and transform that data before analytics delivery — the documented Datastream-plus-Dataflow pattern.
What changed
Event Pipeline from Pub/Sub Through Dataflow to BigQuery
Exam tip: Choose a direct Pub/Sub BigQuery subscription with a UDF for simple delivery to BigQuery, and choose a Dataflow pipeline when the stream needs complex transforms, joins, or enrichment before loading.
Domain 3 of 3
137 practice questions
Practice this DomainLast updated