GCP Professional Cloud Architect Practice Question

A retail company has two managed instance groups (MIGs) in Google Cloud.

One MIG hosts the customer-facing e-commerce site behind a global external HTTP(S) load balancer and must maintain 99.9% availability during flash-sale spikes.
The second MIG runs ad-hoc analytics jobs that checkpoint every minute and can restart without impact if interrupted. The architecture team must cut compute costs while meeting both workload requirements. Which instance type combination should they use for the two MIGs?

Use Spot VMs for both MIGs and rely on load-balancer health checks and automatic restarts to handle preemptions.
Run the e-commerce MIG on standard VMs to ensure availability, and run the analytics MIG on Spot VMs with autoscaling sized to finish on time.
Run both MIGs on standard VMs and purchase one-year committed use discounts to reduce costs.
Run the e-commerce MIG entirely on Spot VMs for lower cost, and run the analytics MIG on standard VMs; enable autoscaling on both groups.

Report Issue

Answer Description

Spot VMs (formerly preemptible VMs) cost roughly 60-91% less than standard VMs because Google can reclaim them at any time and will forcibly stop them after 24 hours. They are therefore ideal for short-lived, fault-tolerant batch or analytics workloads that can resume from checkpoints. In contrast, a customer-facing e-commerce service with a 99.9% availability target needs stable capacity; running it on Spot VMs would risk unplanned shutdowns that violate the SLA. Deploying the e-commerce MIG on standard VMs (optionally with autoscaling or committed-use discounts) and the analytics MIG on Spot VMs with autoscaling minimizes cost while preserving required reliability.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.