The maximum amount of traffic or demand a system experiences during its busiest period. Like a highway during rush hour — everyone is trying to use the road at the same time, and without enough lanes, things slow to a crawl. Cloud systems use auto scaling to handle peak loads without crashing.
An online retailer's website normally handles 1,000 visitors per minute, but during Black Friday sales it spikes to 50,000 visitors per minute. AWS Auto Scaling automatically launches additional EC2 instances to handle the peak load, then scales back down once the rush is over — so the site stays fast without paying for extra servers year-round.
Peak load is a demand pattern, not a single product. These autoscaling services add/remove compute capacity automatically to maintain performance during traffic spikes.