How do you implement auto-scaling policies to handle sudden spikes in traffic without over-provisioning resources?
