Explanation:
The Horizontal Pod Autoscaler (HPA) based on an external metric (option C) is the correct choice. The scenario describes a workload where scaling needs to respond to the volume of Pub/Sub messages, which is an external system metric (e.g., Pub/Sub subscription backlog size). HPA can scale the number of pods based on this metric to handle varying message rates. Vertical Pod Autoscaler (options A/B) adjusts CPU/memory per pod, which is irrelevant here since the problem states pods handle a fixed message rate. HPA based on resource utilization (option D) would not directly align with the message backlog, making it less effective for this use case.
Ultimate access to all questions.
What GKE feature should you use to automatically scale your Deployment to handle varying message processing demands from Pub/Sub, including occasional large batches, while maintaining timely processing?
A
Vertical Pod Autoscaler in Auto mode
B
Vertical Pod Autoscaler in Recommendation mode
C
Horizontal Pod Autoscaler based on an external metric
D
Horizontal Pod Autoscaler based on resources utilization
No comments yet.