A TPM's team is building a system that must handle 10,000 requests per second at peak. Which capacity planning approach is most appropriate?
-
A
Provision for average load and add capacity reactively when alerts fire
-
B
Model peak load with a safety margin, use load testing to validate, and define auto-scaling policies
-
C
Overprovision by 10x to eliminate any risk of overload
-
D
Rely entirely on the cloud provider's default autoscaling without load testing