Autoscaling

Autoscaling automatically adjusts compute capacity to match demand — adding instances when traffic spikes and removing them when it drops. Combined with per-second billing, it lets teams pay only for the capacity actually used while keeping latency within target during peaks.