Associate Data Practitioner

By Google · Question Mode

QuestionQ33

Data Pipeline Orchestration

An organization sends IoT event data to a Pub/Sub topic. Subscriber applications read the messages and transform them before storing them in the data warehouse. During especially busy periods, when more data is written to the topic, the subscriber applications do not acknowledge messages before the deadline. You need to change the pipeline to handle these activity spikes and continue processing messages. What should you do?

Explanation

Subscriber flow control applies backpressure by limiting outstanding messages or bytes delivered to each subscriber. Once those limits are reached, the subscriber stops pulling more messages until in-flight messages are acknowledged or negatively acknowledged, preventing overload during transient traffic spikes.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!