Fastly, one of the largest content distribution networks (CDNs) on the Internet, crashed on Tuesday, causing millions of users to be unable to access certain websites. 85% of the network covered by Fastly returned bugs Tuesday morning and was all caused by a single-user configuration update that discovered a bug that had been dormant in Fastly's infrastructure since mid-May .
In an interview with The Guardian, Fastly’s head of infrastructure and engineering, Nick Rockwell, explained what had really happened to bring down his services and also apologized for the disruption. It’s weird that such a large company is so transparent about it, but it’s certainly a welcome thing.
Content delivery networks operate on the principle that the Internet is faster and more stable if users are physically closer to them. This translates into faster downloads, better security and many other features.
It also means that there is a point of failure if something goes wrong, and that is exactly what happened on June 8th.
"On May 12, we began a software deployment that introduced an error that could be triggered by a specific client configuration in specific circumstances." Rockwell told The Guardian. "In early June 8, a client introduced a valid configuration change that included the specific circumstances that caused the error, which caused 85% of our network to return errors."
Essentially, an error had been introduced to the system on May 12, but it had remained dormant until a client updated the configuration on June 8, which caused the error and caused most of the Internet ( including PC Gamer) for many users. He quickly spotted the problem in a minute and "in 49 minutes, 95% of our network was working normally."
There is a financial impact to preventing users from accessing the sites, of course, and SEO agency Reboot estimates that downtime cost Amazon $ 32 million in sales. Gulp.