Cloudflare down: Why half the internet stopped working on Tuesday

If you spent half of Tuesday morning turning your Wi-Fi off and on again, you weren’t the only one.

Cloudflare, a company that quietly props up a massive chunk of the internet, had a major wobble earlier today. It is one of those background services you probably never think about until it breaks. But when it does break, it tends to take the rest of the digital world down with it.

It all kicked off around 11:30 AM UK Time. Reports started flooding in that some of the world’s biggest websites were failing to load. If you tried to check your feed on X (formerly Twitter), finish a presentation on Canva, or ask ChatGPT a question, you likely hit a brick wall.

Instead of the usual content, users were met with cryptic error messages. Some saw a generic ‘Internal Server Error 500’, while others were presented with a broken page asking them to ‘unblock challenges.cloudflare.com’. It wasn’t just frustrating; it was confusing. The websites themselves weren’t technically down, but the door to get into them was jammed shut.

The invisible middleman

To understand why this happened, you have to look at what Cloudflare actually does.

In the old days of the web, your computer would connect directly to a website’s server. That was simple, but it meant those servers could easily get overwhelmed by too many visitors or malicious attacks.

Cloudflare sits in the middle. It acts like a bouncer and a traffic warden rolled into one. It sits between you and the website, filtering out bad traffic and making pages load faster by serving them from data centres closer to your home.

Most of the time, this makes the web faster and safer. But it also creates a massive bottleneck. Because so many different companies – from news sites to gaming platforms – rely on Cloudflare’s tools, a single glitch at their end can knock out a huge variety of seemingly unconnected services all at once.

Who was hit by the outage?

The outage was surprisingly widespread. It wasn’t just social media and work tools. Gamers trying to log into League of Legends or Valorant found themselves unable to connect to servers.

In a bit of ironic bad luck, even Downdetector – the site everyone rushes to when they think the internet is broken – went offline briefly. That is because they, too, rely on Cloudflare to keep their site running. When the tool you use to check for outages is having an outage, you know it’s a bad morning.

A fragile web infrastructure

This incident is a bit of a wake-up call about how the modern internet works. We are moving toward a world where the entire web relies on a handful of massive infrastructure giants.

This isn’t an isolated event, either. Just last month, Amazon Web Services (AWS) suffered a similar outage that knocked more than 1,000 sites and apps offline. Microsoft’s Azure platform had its own issues shortly after.

It highlights a certain fragility in the system. When one of these infrastructure providers sneezes, the rest of the internet catches a cold.

Is it fixed?

Thankfully, yes. By around 12:20 UK Time, things started to stabilise. Cloudflare confirmed they were ‘aware of, and investigating’ the issue. They haven’t released a full technical breakdown of what went wrong yet, though some keen-eyed observers noted that there was scheduled maintenance happening in their Santiago data centre right around the time the trouble started.

For now, services seem to be recovering, though Cloudflare has warned that customers might still see ‘higher-than-normal error rates’ while they finish mopping up the mess.

So, if your favourite site is still loading a bit slowly this afternoon, give it a few minutes. It’s not your router, and it’s not your phone. It’s just the plumbing of the internet getting a much-needed fix.