Amazon’s cloud computing unit Amazon Web Services experienced an outage on Tuesday, affecting publishers that suddenly found themselves unable to operate their sites.

The company said on its website that the root cause of the issue was tied to a service called AWS Lambda, which lets customers run code for different types of applications.

Roughly two hours after customer began experiencing errors, the company posted on its AWS status page that many of the affected AWS services were “fully recovered” and it was continuing to recover the rest.

Amazon said it had experienced multiple error rates for AWS services in the Northern Virginia region where it clusters data centers. The company said customers may be dealing with authentication or sign-in errors when using some AWS services, and experiencing challenges when attempting to connect with AWS Support. The issue with Lambda also indirectly affected other AWS services

Patrick Neighorn, a company spokesperson, declined to provide additional details about the outage.

The outage was first confirmed shortly after 3 p.m. ET. and it was unclear how widespread the problem extended. But many companies, including news organizations such as The Verge and Penn Live, said they were experiencing issues. The Associated Press was also hampered by the outage, unable to operate their sites amid breaking news that former President Donald Trump was appearing in court in Miami.

The episode on Tuesday is reminiscent of a much longer AWS outage in December 2021, which affected a host of U.S. companies for more than five hours.

The outage comes as Amazon is holding a two-day security conference in Anaheim, California to tout its cloud offerings to its clients or other companies that might be interested in storing their data on its vast network of servers around the world. Companies have been cutting back their spending on the unit, leading to slowdown in growth during the most recent quarter.