Question

A financial services company is deploying a new stateless REST API that experiences highly variable traffic, spiking from 200 to 8,000 requests per second during market open. The business requires that 95th-percentile response latency stay under 300ms during spikes, that the service scale automatically without manual intervention, and that infrastructure costs remain minimal during off-peak hours when traffic is near zero. Which compute design best meets these technical requirements?

Accepted Answer

Deploy the API on Cloud Run with concurrency tuning and autoscaling configured with a minimum of 0 instances and a maximum sized for peak load.

Answer

Deploy the API on a fixed fleet of Compute Engine instances sized for peak traffic behind an external HTTP(S) load balancer.

Answer

Deploy the API on a Compute Engine managed instance group with autoscaling based on average CPU utilization at 60% and a minimum of 10 instances.

Answer

Deploy the API on a single large Compute Engine instance with a vertical autoscaling policy that resizes the machine type based on load.