Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Developer practice bank has an additional 608 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
Your team is designing a stateless, fault-tolerant document processing service on Compute Engine. The application experiences a steady baseline load 24/7 along with unpredictable, short-duration traffic spikes throughout the week.
You need to design a compute and scaling strategy that:
Which compute and scaling strategy should you implement?
ANY or BALANCED distribution shape, and an instance flexibility policy using smaller machine types across multiple families.This strategy separates predictable baseline capacity from unpredictable burst capacity by combining Compute Engine reservations and resource-based Committed Use Discounts (CUDs) with an autoscaling regional Managed Instance Group (MIG) powered by Spot VMs and infrastructure flexibility.
ANY or BALANCED target distribution shape provides location flexibility across multiple zones. Pairing this with an instance flexibility policy across multiple machine families (for example, combining N4, N2, and E2 series) and smaller machine types ensures Compute Engine can provision available fallback hardware if a primary machine type or zone lacks capacity or preempts a Spot VM.Unlike over-provisioning reservations for peak load or relying on a single machine type in one zone, this architecture aligns billing commitments with guaranteed baseline usage while maximizing elasticity and cost savings for variable demand.
regionInstances.bulkInsert API with an instance flexibility policy, and rely on Compute Engine rightsizing recommendations to adjust machine sizes when traffic fluctuates.bulkInsert and rightsizing recommendations doThe regionInstances.bulkInsert API creates a batch of standalone Compute Engine VMs in a single request across zones in a region. Rightsizing recommendations analyze Cloud Monitoring CPU and memory telemetry over the previous 8 days to suggest resizing overprovisioned or underprovisioned standalone VMs.
bulkInsert sets minCount equal to count, meaning the entire request fails if all requested VMs cannot be provisioned upfront.You needed a one-time bulk creation of standalone VMs for a fixed-size task and wanted to periodically optimize static VMs showing steady, multi-week overprovisioning.
To automatically add, remove, and repair instances across multiple zones and machine families as demand changes, workloads should use an autoscaling regional MIG rather than standalone instances.
Zonal reservations guarantee physical capacity for a specific machine type in a single zone, while 3-year resource-based Committed Use Discounts (CUDs) provide maximum discounts in exchange for a continuous three-year commitment to specific hardware.
The workload required strictly fixed hardware running at a constant, predictable peak utilization 24/7 with deterministic single-zone placement and zero tolerance for hardware variance or preemption.
When workloads exhibit a steady baseline paired with unpredictable spikes, best practices recommend reserving only the minimum predictable baseline and covering variable peaks with flexible, elastic compute options such as Spot VMs or compute flexible CUDs.
The Flex-start provisioning model uses Dynamic Workload Scheduler (DWS) to queue requests for high-demand resources (such as GPUs, TPUs, and H4D instances) for up to seven days, provisioning them at a discount once capacity becomes available.
The workload were a time-flexible batch or AI/ML job—such as model fine-tuning, HPC simulation, or batch inference—that requires scarce GPU or TPU hardware for up to seven days and can tolerate waiting in a queue to start.
Infrastructure flexibility allows combining location flexibility with either machine type flexibility or time flexibility, but machine type flexibility and time flexibility are mutually exclusive.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated