Eugene Svboda, a Metro reader, attempted to log onto his bank just after 11am yesterday.
Yet after typing his username and password, Eugene was instead greeted by a grey screen that showed him a word he had never heard of before.
‘Cloudflare,’ Eugene tells Metro. ‘First I thought it was malware. Used a second laptop at around 1pm and got the same message from Cloudflare.’
Eugene never got into his online banking that day – but he was not the only person shown a certain grey error screen.
Cloudflare, a company that helps websites secure and manage their internet traffic, experienced an issue with its global network yesterday.
Users began reporting problems with websites and apps that use Cloudflare at around 11am, according to internet outage tracker Downdetector, which itself was briefly knocked offline.
Within minutes, websites were offline for as long as four hours, including X, ChatGPT, PayPal and League of Legends.
What caused the Cloudflare outage?
Cloudflare is the Swiss Army Knife of the tech world. The network provides tools to help websites fight off cyber attacks and load content.
One safety check it does is confirming whether a user trying to load a website is a human or a bot.
Technical jargon aside, Cloudflare verifies a user’s, well, humanity by writing up a ‘file’.
But when engineers changed the way the system generates this file, the system instead duplicated this file so many times that it crashed.
Luke Kehoe of Ookla, the company behind Downdetector, tells Metro that yesterday’s outage impacted 3.3 million people.
‘The US accounted for roughly a quarter of these (823,000+ reports), with the UK second at 284,000+, followed by Germany (214,000+), Brazil (200,000+), Japan, the Netherlands, Poland, France, Canada and Italy,’ the industry analyst says.
Are we too reliant on services provided by a few giant companies like Cloudflare?
Simply put, yes, every tech expert Metro spoke with said. We’ve seen three major outages in only a few weeks.
Last month, an issue with a data centre in Northern Virginia forced 2,000 websites and apps offline for more than two hours.
The data centre is owned by Amazon Web Services, the e-commerce giant’s cloud service provider, relied upon by millions of firms, including Reddit, Snapchat, Netflix and various governments.
Days later, Azure, Microsoft’s cloud service system, went offline. This disrupted airlines, mobile networks and supermarkets for eight hours.
Cloudflare’s outage similarly hit hard because 20% of the web runs through its network.
For yet another outage to occur is a reminder that certain companies have an outsized role in keeping the world online, says software artefact management firm Cloudsmith’s CTO, Lee Skillen.
‘Today’s Cloudflare, tomorrow might be Fastly,’ he explains to Metro, referring to the US cloud platform used by 1,200 companies.
‘Although outages are not uncommon, a global “completely down and out” outage like this is absolutely highly unusual, and there is no doubt that this has a wide-reaching impact worldwide for businesses and their users.’
Benjamin Schilz, CEO of the digital workspace platform Wire, agrees. ‘Modern society is built on the assumption that connectivity never fails,’ he says to Metro.
‘The problem isn’t the Cloudflare outage itself. It’s the brutal dependency we’ve created: a handful of global providers carrying the weight of an entire digital economy.
‘If the internet falters, life as we know it halts.’
Could the entire internet be taken offline?
Thankfully, there’s no off switch to the internet, which consists of countless tiny bits of code zipping around wires as thin as hair strung across the ocean floor.
Small blackouts do occur, especially in regions ravaged by war. Only earlier today did this happen in Ukraine after it was pummelled by Russian forces, internet analytics firm NetBlocks said.
An ‘internet kill switch’ has also been used by governments in China, Egypt and Iran to quell protests by forcing internet providers to shut down or erecting powerful firewalls.
The UK can push the big red button in the event of a national crisis, such as a major cyber attack, but this power has never been used.
But the internet is a big web – a World Wide Web – made up of the networks that governments, companies and people run.
Many websites don’t use the services offered by Amazon and other tech giants. Savvy internet users have always managed to get around government blackouts.
And, even when an earthquake near Taiwan damaged critical communication cables in 2006, a good chunk of the world remained online.
So the only reason the entire internet could collapse would be the destruction of the world on such a massive scale that you not being able to watch Netflix would be the least of your worries, says Dr Stilianos Vidalis, the deputy head of computer science at the University of Hertfordshire.
‘In the 21st century, society and companies rely on the Internet for functioning and for offering their services,’ he tells Metro. ‘This reliance can lead to loss of capabilities, loss of resilience and increased dependency on a handful of providers that essentially run the Internet.’
We came close-ish to the ‘net going under in 1988, Dr Vidalis says, when the ‘Morris worm’ disabled 10% of all internet-connected systems.
The worm was a self-replicating programme that exploited known weaknesses in how utilities sent emails and logged people on.
‘I argue that a Morris 2.0 worm would have a rather different impact in 2026,’ Dr Vidalis says.
‘Communications would collapse. We would not be able to use our mobile phones, communication apps, email and indeed any authentication services. Financial transactions would halt and the only available method of payment would be cash.’
Public transport would be disrupted, tech giants like Google and Microsoft would become obsolete and there would be an information blackout.
‘The markets would start crashing almost immediately as governments would start realising that the globalised economy cannot function,’Dr Vidalis adds.
‘This catastrophic scenario is highly unlikely as the Internet has built-in resilience.’
We might not have another Y2K in the near future, but that doesn’t mean the internet isn’t fragile, says Kashif Nazir, a senior technical architect at the app migration company Cloudhouse.
‘The internet was designed to survive nuclear war, but we’ve essentially re-centralised it and handed the keys to five companies,’ he tells Metro.
‘When Cloudflare goes down, it doesn’t matter that the underlying infrastructure is fine; for millions of users, the internet is effectively down.’
Get in touch with our news team by emailing us at webnews@metro.co.uk.
For more stories like this, check our news page.