News & Politics

'We Let the Internet Down Today': Cloudflare CEO Apologizes for 'Bug' That Took Down Millions of Websites

(Getty Images)

Internet network and security provider Cloudflare experienced a massive outage this morning, causing thousands and maybe millions of websites to shut down briefly.

Websites in Europe and eastern North America were the worst hit because of general European and American internet activity around those times.

Cloudflare is an internet-hosting platform that many internet services rely on to be functional and protected in case of cyberattacks. Around 10 percent of all websites, including ones like Drudge Report, Medium, Dropbox, Pinterest, and Discord, use this service.

Several cryptocurrency trading groups, such as Coinbase and CoinMarketCap, also depend on Cloudflare and incorrectly displayed the price of bitcoin during the outage.

At 10:08 a.m., Cloudflare co-founder and CEO Matthew Prince tweeted, “Aware of major @Cloudflare issues impacting us network wide. Team is working on getting to the bottom of what’s going on. Will continue to update.”

After around 30 minutes, Cloudflare was able to implement a fix for the issue. The company states, “For about 30 minutes today, visitors to Cloudflare sites received 502 errors caused by a massive spike in CPU [central processing unit] utilization on our network. This CPU spike was caused by a bad software deploy that was rolled back. Once rolled back the service returned to normal operation and all domains using Cloudflare returned to normal traffic levels.”

“This was not an attack (as some have speculated) and we are incredibly sorry that this incident occurred,” the company adds.

Cloudflare experienced internet outages as recently as June 24, 2019. On that day, many websites on Cloudflare were unavailable because a small company in Northern Pennsylvania became the preferred path of too many internet users.

Prince indicates that the outage today is of a different nature.

The previous incident occurred due to issues with the Border Gateway Protocol, a fundamental internet routing system. However, today’s glitch stemmed from a “bug on our side that caused Firewall process to consume excessive CPU.”

Prince adds that Cloudflare is “putting in place systems so never happens again.” But, he still tells Bloomberg in an interview that “we let the internet down today.”