An enterprise data engineering team processes nightly batch data pipelines that are fault-tolerant and save progress checkpoints to a Cloud Storage bucket every 10 minutes. The engineering manager needs to optimize compute costs for this workload while delegating operational tasks—such as starting, stopping, and reconfiguring these Compute Engine VM instances—to a specialized operations team following Google Cloud security best practices. Which TWO actions should the cloud engineer take to meet these requirements? (Select 2 answers.)
- Provision Compute Engine virtual machines using the Spot VM provisioning model to run the batch processing tasks.Answer
- Grant the operations team the predefined Compute Instance Admin (v1) role (roles/compute.instanceAdmin.v1) on the project.Answer
- CProvision standard persistent On-Demand VM instances because Spot VMs cannot write persistent checkpoint data to Cloud Storage buckets.
- DAssign the primitive Editor role (roles/editor) to the operations team to ensure they have sufficient permissions to modify instance settings.
- EAutomate the creation of a new Google Cloud project when region CPU capacity limits are met to bypass quota restrictions without manual intervention.
Answer
The cloud engineer should provision Compute Engine virtual machines using the Spot VM provisioning model and grant the operations team the predefined Compute Instance Admin (v1) role (roles/compute.instanceAdmin.v1).
Utilizing Spot VMs is Google Cloud's recommended strategy for batch and fault-tolerant workloads that checkpoint progress, yielding maximum cost efficiency. Additionally, delegating VM management via the predefined `roles/compute.instanceAdmin.v1` role adheres strictly to the principle of least privilege while providing all required Compute Engine operational capabilities.
Step-by-Step Solution
Key Concept
Managing Compute Engine resource cost optimization with Spot VMs and securing lifecycle access using IAM predefined roles.