Google Professional Cloud Developer Question 412
Select 2Google Cloud PlatformYou are developing an application that processes real-time data from a Google Cloud Pub/Sub topic. The application must scale dynamically based on the volume of messages while ensuring at-least-once delivery. Which of the following approaches should you implement to meet these requirements?
- A
Use a pull subscription with a message acknowledgment mechanism and deploy the application on a Google Kubernetes Engine (GKE) cluster with horizontal pod autoscaling enabled.
- B
Use a push subscription to send messages directly to a Cloud Function, which processes the messages and ensures acknowledgment.
- C
Use a pull subscription and deploy the application on a Compute Engine instance with a fixed number of instances to ensure predictable scaling.
- D
Use a Dataflow streaming pipeline to process messages from the Pub/Sub topic and enable autoscaling for the Dataflow job.
Show answer and explanation
Correct answers: A, D
Explanation
To meet the requirements of dynamic scaling and at-least-once delivery, choosing a pull subscription with GKE autoscaling or a Dataflow streaming pipeline are the best options. These solutions allow the application to handle varying message volumes effectively while maintaining reliable processing. Push subscriptions or fixed Compute Engine instances lack the necessary scalability or flexibility in this scenario.
- A. Correct.
This option is correct because a pull subscription allows the application to control the acknowledgment of messages, ensuring at-least-once delivery. Deploying on GKE with horizontal pod autoscaling allows the application to scale dynamically based on the message volume.
- B. Incorrect.
This option is partially correct for ensuring at-least-once delivery but does not allow dynamic scaling based on message volume since Cloud Functions scale based on the number of concurrent executions, which may not handle all scenarios effectively.
- C. Incorrect.
This option is incorrect because using a fixed number of Compute Engine instances does not allow for dynamic scaling based on message volume. This contradicts the requirement for scalability.
- D. Correct.
This option is correct because Dataflow natively integrates with Pub/Sub, supports at-least-once delivery, and offers autoscaling for streaming pipelines, making it a suitable choice for this scenario.