Every system works perfectly... until people start using it

One thing I've noticed over the years is that most systems don't fail because of bad code.

They fail because they were never expected to handle what came next. When an application has 10 users, almost any architecture looks good.

• At 1,000 users, you start seeing delays.
• At 10,000 users, bottlenecks become visible.
• At 100,000 users, those bottlenecks become everyone's problem.

That's why I think the first lesson in System Design is surprisingly simple: Every system has a bottleneck.

The challenge isn't avoiding bottlenecks. The challenge is knowing where they're likely to appear. It could be:

• a database handling every request
• a single service doing too much work
• an API that wasn't built for the traffic it's receiving

The mistake many teams make is assuming they'll figure out scaling when growth arrives.

By the time growth arrives, you're often making architectural decisions under pressure.

The engineers I learn the most from don't obsess over designing for millions of users on Day 1.

They simply ask:If this system grows 10x, what breaks first?

That's usually where the real System Design conversation begins.

Question: What's the first bottleneck you've seen become a production issue?