Cloudflare Faces Second Service Disruption in Weeks, Raising Questions About Internet Centralisation: If you tried opening LinkedIn, Zoom, Canva, or even Downdetector early Friday morning and were met with error messages, you weren’t alone. For about half an hour, a surprising chunk of the internet simply… blinked out. The culprit? Cloudflare—again.
This marked the company’s second outage in a matter of weeks, and while Friday’s disruption was brief and relatively small compared to last month’s chaos, it highlights something much bigger: our digital world may be a lot more fragile than we’d like to admit.
What Actually Happened This Time?
Cloudflare explained that the issue wasn’t an attack or a failure of hardware. Instead, it came as the unintended side effect of a security tweak. Earlier in the week, a major software vulnerability became public, prompting Cloudflare to update its firewall rules to protect customers. That change, meant to keep servers safe, ended up temporarily making some of them unreachable.
For about thirty minutes, Cloudflare’s systems wobbled, and so did every site leaning on its infrastructure. By a little after 9am GMT, the issue was resolved, but the ripple effects—frustrated users, confused companies, and thousands of error reports—lingered online.
A Pattern Is Starting to Form
Friday’s hiccup follows a much wider outage in mid-November, one that disrupted platforms used by hundreds of millions of people. Sites like X, OpenAI, Spotify, and popular online games all went down, and the explanation was decidedly unglamorous: an auto-generated configuration file grew too large, breaking the software responsible for managing traffic.
In other words—one file got too big, and major chunks of the internet ground to a halt.
When outages this large happen only weeks apart, it’s natural for people to start asking questions. Is this unlucky timing? Or does Cloudflare’s sprawling infrastructure have deeper systemic issues?
Even experts say it’s too early to tell—but they agree that the pattern is worrying.
Centralisation: Convenience Today, Risk Tomorrow
Cloudflare is popular for a reason. It protects websites from cyberattacks, makes pages load faster, and boosts reliability. For many companies, it’s a no-brainer solution. Why manage complex networking setups when Cloudflare can handle it at global scale?
But with convenience comes concentration, and concentration introduces risk.
University College London professor Steven Murdoch summed it up well: Cloudflare’s rapid growth is a testament to the strength of its products, but it also creates a single point of potential failure. When a piece of Cloudflare’s infrastructure slips—even briefly—the consequences go far beyond a single website. Suddenly, millions of people around the world feel it.
And Cloudflare isn’t alone in this trend. Just last month, Amazon Web Services experienced an outage that affected more than 2,000 businesses worldwide. When a handful of companies provide the backbone for so much of the internet, any disruption—technical, accidental, or malicious—spreads fast.
Experts Are Ringing the Alarm Bell
Michał “rysiek” Woźniak, an expert in DNS and global internet infrastructure, put the situation bluntly: the modern internet built by big tech companies is brittle. Friday’s interruption marks the fourth major outage since late October—four incidents big enough for non-technical people to notice and care about.
This is not the resilient, decentralised network that early internet pioneers envisioned. Instead, it’s a tightly interconnected system in which everyday digital life depends on a few players—Cloudflare, Amazon, Google, Microsoft—working perfectly, all the time.
So… What Now?
None of this is a call to abandon Cloudflare or giant cloud services. They do offer unmatched security, speed, and reliability for millions of customers. But these back-to-back outages show that the internet’s foundations might need more redundancy, more distribution, and more serious conversations about infrastructure resilience.
If the internet is going to keep growing in complexity—and our dependence on it keeps increasing—we can’t afford for a single misconfigured file or a security tweak to take the digital world offline.
For now, Friday’s outage is over. But the broader debate it sparked is only just beginning.
Reimagining Visual Generation with Nano Banana Pro (Gemini 3 Pro Image) | Maya
