Large parts of the internet appear to have broken, seemingly because of an issue at Google.
Google's own services such as Gmail, Google Drive, its Nest smart home service and just about every product in its Gsuite set of apps were broken by the problems with the company's own cloud services. But it also hit sites that are not owned by Google but rely on its technology, including Uber, Vimeo, and chat app Discord.
Users complained it was difficult or impossible to log in to the sites, which rely on Google's cloud services to power their central functions.
Together, the sites represent a significant proportion of the internet as people use it. On YouTube alone, people watch more than a billion hours of video each day.
The issues lasted for more than an hour before Google claimed to have located the issue and said it had been fixed.
"We are experiencing high levels of network congestion in the eastern USA, affecting multiple services in Google Cloud, G Suite and YouTube," a note on Google's page reads. "Users may see slow performance or intermittent errors. We believe we have identified the root cause of the congestion and expect to return to normal service shortly."
While outage maps such as those at tracking website Downdetector suggested the problems were particularly focused on the east coast of the US, issues were being reported across the world.
Earlier, the official YouTube site had acknowledged the issue and said it was being worked on.
"If YouTube isn't loading for you or you're experiencing error messages, we're working to fix it!" it posted.
Similarly, Snapchat said it was aware of the problem and that it was being worked on.
"We're aware of an issue preventing some Snapchatters from using the app," it wrote, before posting a stressed-looking emoji. "Hang tight!"
Though it might not be obvious to users of the apps or services themselves, much of the underlying infrastructure of the web is provided by just a few companies. That means that issues at one of those firms – such as Google or Amazon – can very quickly cause problems across the internet, taking down websites that might appear entirely separate from those suffering the problems.
In 2017, for instance, large parts of the internet stopped working because of an issue at Amazon Web Services, which caused problems at Quora, Trello and some of the world's biggest news sites. That was later traced back to one small typo in the company's code, which caused underlying infrastructure to break and websites to stop working.
Similarly, in 2016, a vast cyber attack focused on Dyn, which also provides some of the infrastructure that keeps the web breaking. That meant websites like Spotify, Reddit and Twitter all went down.
Join our commenting forum
Join thought-provoking conversations, follow other Independent readers and see their replies