Google Professional Cloud DevOps Engineer Question 170
Single answerGoogle Cloud PlatformYour team has deployed a high-traffic, machine learning application on Google Cloud using several Compute Engine instances in an autoscaled managed instance group. The application is currently experiencing higher than expected costs. Upon investigation, you notice that the CPU utilization of the instances rarely exceeds 30%, and the application’s latency is within acceptable limits. What should you do to optimize resource utilization and reduce costs?
- A
Reduce the size of the instance machine type used by the managed instance group.
- B
Disable autoscaling for the managed instance group to prevent scaling beyond the current instance count.
- C
Configure a custom image for the instances to improve application performance.
- D
Enable preemptible VM instances in the managed instance group to replace standard instances.
Show answer and explanation
Correct answer: A
Explanation
The low CPU utilization indicates that the current machine type is overprovisioned. By reducing the size of the instance machine type, you can better match resource allocation to the workload demands, optimizing utilization and lowering costs without negatively impacting latency or performance.
- A. Correct.
Reducing the instance machine type size aligns with the observed low CPU utilization, optimizing resource usage and reducing costs without affecting application performance.
- B. Incorrect.
Disabling autoscaling could result in insufficient resources during traffic spikes, leading to application downtime or performance degradation, and does not directly address resource underutilization.
- C. Incorrect.
Configuring a custom image may improve startup time or specific operations but does not address the low CPU utilization or directly reduce costs.
- D. Incorrect.
Using preemptible VMs could reduce costs but is not suitable for applications that require high availability, like this high-traffic application.