Marcio Jose Sanchez / AP
LONDON – Quickly, the company hit by a major outage that took many of the world’s best websites offline this week blamed the problem on a software bug that triggered when a customer made changes a parameter.
The problem at Fastly meant internet users couldn’t log into a slew of popular websites early on Tuesday, including the New York Times, the Guardian, Twitch, Reddit and the UK government homepage.
“We experienced a global outage due to an undiscovered software bug that emerged on June 8 when it was triggered by a valid customer configuration change,” said Nick Rockwell, senior vice president of engineering and infrastructure at Fastly, in a blog post Tuesday night.
He said the outage was “large and severe,” but the company quickly identified, isolated and deactivated the problem and after 49 minutes most of its network was back up and running. The bug was included in a software update that was rolled out in May and Rockwell said the company is trying to figure out why it was not detected during testing.
“Even though there were specific conditions that triggered this outage, we should have anticipated it,” said Rockwell.
San Francisco-based Fastly provides what’s called a content delivery network, an arrangement that allows clients’ websites to store data such as images and videos on various mirror servers in 26 countries. Keeping data closer to users means it is displayed faster.
But the incident highlighted how much of the global internet depends on a handful of backroom companies like Fastly that provide vital infrastructure, and it amplified concerns about their vulnerability to disruption. more serious.