Examkin

Question

Difficulty: MediumManaging Compute Engine Resources

An infrastructure team manages a workload running on Compute Engine virtual machines that requires autoscaling based on system memory utilization. Currently, Cloud Monitoring only displays default hypervisor-level metrics, which do not include internal RAM usage. Which action should the team take to collect guest memory metrics and enable autoscaling for the instance group?

  1. Install the Google Cloud Ops Agent on the virtual machines, and configure the autoscaling policy using the memory metrics exported to Cloud Monitoring.Answer
  2. B
    Install the legacy Stackdriver Monitoring Agent on the virtual machines to enable hypervisor-level memory reporting without configuring custom metrics.
  3. C
    Grant the Compute Admin primitive role to the VM default service account so Cloud Monitoring can inspect guest memory directly.
  4. D
    Convert the instances to Spot VMs, which natively expose guest memory metrics to hypervisor monitoring without additional agent software.

Answer

Install the Google Cloud Ops Agent on the virtual machines, and configure the autoscaling policy using the memory metrics exported to Cloud Monitoring.
Hypervisor metrics natively collected by Compute Engine do not include internal operating system details like memory usage. To collect RAM utilization, the Google Cloud Ops Agent must be installed inside the guest OS. This agent streams telemetry to Cloud Monitoring, allowing the Managed Instance Group (MIG) autoscaler to scale based on memory utilization.

Step-by-Step Solution

1
Analyze metric availability
Determine that memory utilization is an internal guest OS metric not captured by default hypervisor metrics.
Compute Engine hypervisors only capture host-level metrics such as CPU utilization and disk I/O.
2
Select the appropriate telemetry agent
Deploy the Google Cloud Ops Agent onto the virtual machine instances.
The Google Cloud Ops Agent is the Google-recommended solution for gathering guest system metrics and logs.
3
Configure Managed Instance Group autoscaling
Set the autoscaling policy target metric to the Ops Agent memory utilization metric in Cloud Monitoring.
MIG autoscalers can scale based on custom metrics or Ops Agent telemetry ingested into Cloud Monitoring.

Key Concept

Managing Compute Engine Resources with Telemetry and Autoscaling
Rate this question

Topics

Managing Compute Engine Resources

More questions from this topic