Major leading websites and services across the United States and around the world were recovering early Monday after a problem at a major cloud computing service left leading games, publishers and streaming platforms unusable to millions. 

Issues on some apps and websites, including Snapchat, Roblox, streaming site Max and PayPal’s Venmo were showing signs of easing, according to Downdetector.

Other services, however, remained affected, with thousands of reports for social media app Reddit and financial platform Chime on Downdetector.

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

The outage underlined the fragility of companies — including financial services — that use cloud-based servers to host their data, and how suddenly businesses across the globe can be impacted by an unplanned outage. 

AWS first reported a problem at 12:11 a.m. PT (3:11 a.m. ET) and said it was dealing with an “operational issue” affecting 14 different services in its center in northern Virginia, at its U.S.-East-1 Region center.

At 6:35 a.m ET, the company said the database problem that caused the outage was “fully mitigated” but but warned that there may still be delays.

Ookla, owner of outage tracking website Downdetector, said over 4 million users reported issues due to the incident.

”These disruptions are not just technical issues, they’re democratic failures. When a single provider goes dark, critical services go offline with it — media outlets become inaccessible, secure communication apps like Signal stop functioning, and the infrastructure that serves our digital society crumbles,” said Corinne Cath-Speth, Article 19’s head of digital.

AWS customers were unable to report the problem because its automated support ticking system was also offline.