A massive internet outage affected multiple sites and services around the world on Tuesday. Users saw statements like "503 error" or "connection lost" when they've tried to access some of their favourite sites online. 

Popular websites including Reddit, Spotify, Twitch, Stack Overflow, GitHub, gov.uk, Hulu, HBO Max, Quora, PayPal, Vimeo, Shopify, as well as news outlets like the Guardian, the Financial Times and the BBC were all affected.

The cause of the problem appears to have been a glitch at CDN provider Fastly. A CDN, which stands for Content Distribution Network, enables websites to load considerably faster, as well as protect them from attacks plus deal with unpredictable bursts of traffic. 

Fastly, the CDN provider in question confirmed on its status page that it was responsible for the outage that started around 11am UK time (10am UTC) on Tuesday 8 June. 

A statement from Fastly at 10.57am UTC later stated that: "The issue has been identified and a fix has been applied. Customers may experience increased origin load as global services return." 

That issue according to the company involved "a service configuration that triggered disruptions across our POPs globally." The company reported that it has "disabled that configuration," enabling it to bring its "global network ... back online."

A PoP is short for "Points of Presence" and are at the heart of the CDN infrastructure allows the service to serve content from globbaly distributed servers that are closer to the end user. 

Some sites managed to bypass the service to restore functionality, however many were offline for over an hour.  

In 2020 a problem with Cloudfare, another CDN, led to a major internet outage in America and Europe. 

There is no suggestion that it was an attack or foul play, however a number of people turned to Twitter and other social media platforms to question why it was so easy to knock out large parts of the internet around the globe, including the UK government's website. 

With more and more emphasis on doing things online, many services and sites will now be questioning their uptime reliance and whether they have backup plans in place for next time.