When the queue at the supermarket gets long, they open more tills. Websites do the same when lots of people turn up.
A supermarket on a Saturday morning: the queue reaches the back of the shop. A manager presses a button and three more tills open. The queue melts away.
That’s scaling. It means making a system handle more users or more work, usually by adding more computers or making them more powerful.
Two ways to scale
- scaling up: a bigger, more powerful computer, like a faster till
- scaling out: more computers sharing the work, like more tills
Most big websites scale out, because you can keep adding computers almost without limit.
Automatic scaling
The cloud made it possible to add computers in minutes and remove them when it’s quiet, so you only pay for what you need. Serverless does it for you without asking.
What it isn’t
It isn’t only about computers. Sometimes a slow database or a clumsy piece of code is the real bottleneck. Adding machines doesn’t help then.
Where the comparison breaks down
A supermarket can see the queue growing. Websites need monitoring to spot the rush. Sudden spikes can arrive faster than new computers start.
How the words fit together
Work waits in a queue. Scaling adds workers to clear it. The cloud and serverless make that quick.
Big events like ticket sales still crash sites, because demand jumps faster than extra computers can start up. That's why waiting rooms exist.

When you'll hear it
“Will it scale?” Will it keep working if we get ten or a hundred times more users?
“It auto-scales.” It adds computers by itself when it gets busy, then removes them when it's quiet.
“We need to scale up.” We need more capacity to handle the extra demand.
“What happens to our site if a hundred times more people turn up tomorrow?”

