Exam Associate Cloud Engineer topic 1 question 275 discussion

Here's why a Vertical Pod Autoscaler (VPA) is the most suitable solution for this scenario: Right-Sizing Resources: VPA is designed to automatically adjust the resource requests and limits (CPU and memory) for pods based on their actual usage. This ensures that pods have enough resources to run efficiently without being over-provisioned, which can lead to wasted resources and higher costs. Automated Optimization: VPA continuously monitors the resource usage of your pods and recommends optimal settings. You can choose to apply these recommendations automatically or manually, giving you flexibility and control over the process. Microservice-Specific Tuning: By configuring a VPA for each microservice, you can fine-tune the resource allocation for each individual service based on its specific needs and usage patterns. This is more efficient than making blanket changes to the entire cluster or node pool.

It is A Option A: Configure a Vertical Pod Autoscaler for each microservice is the best solution because the VPA automatically adjusts the resource requests and limits for each microservice, ensuring that memory and CPU resources are correctly sized according to their actual usage patterns. This will address your issue of inappropriate resource limits for many of the microservices in your GKE cluster.

As mentioned "resource limits configured for each container in the deployment", so can not vertically increase resources/memory Cluster scale out is not relevant here. With Horizontal scaling, adding more pods ultimately makes sure each service have "right sized memory and CPU"

Vertical is for more specific resources of the individual pods. Horizontal is for creating more copies of the instances (adding more pods).

A seems better