Millions around the world found themselves unable to access popular services thanks to a Domain Name System issue with Amazon Web Services.

Amazon Web Services, a cloud computing service run by Amazon, experienced a significant outage that disrupted numerous websites on Oct. 20.

Disruptions lasted upward of two hours for most services, though some users—mostly in the United States—continued to see problems for over six hours. AWS announced at 5:27 a.m. ET that core issues had been resolved and most services were recovering, but intermittent disruptions persisted into the morning.

Amazon Web Services, a major provider of cloud hosting that underpins much of the web and everyday online tools, went offline because of a problem with one its core database products.

A massive AWS outage Monday that brought down some of the world’s most popular apps and services all started with a glitch.

A major outage caused by an issue at Amazon Web Services affected hundreds of websites, games and apps, including Amazon, Snapchat, Fortnite and several banks. Amazon says it fixed the underlying problem but a full recovery will take longer.

Alaska Airlines already tried to shore up its IT infrastructure after an outage in July forced the Seattle-based company to ground flights across the country.

Monday’s Amazon Web Services outage — and the global disruption it caused — underscored just how reliant the internet has become on a small number of core infrastructure providers.