Amazon on Friday was still wrangling with troubles at its Web hosting service used by Foursquare, Quora, Reddit and other popular websites.
Amazon EC2 was being slowly restored a day after the onset of disruptive problems, but it may take a while to fully recover, according to updates on the service's online status dashboard.
"When a service that is heavily redundant, like this one is, goes out of band this long, they have a serious problem and they don't have their arms around it," said analyst Rob Enderle of Enderle Group in Silicon Valley.
Websites hosted by Amazon EC2 were sluggish or knocked offline on Thursday and the company has yet to disclose a cause.
"These things are supposed to be incredibly bulletproof so it would take something serious to cause a problem like this," Enderle said of datacenters designed to host websites or software in the Internet "cloud."
"If it was something simple, by now they would have told us what it was."
Cloud datacenters such as those at the heart of EC2 are built with redundant systems so loads can be shifted in the event of trouble such as a computer server failing.
A problem disrupting the entire service could range from a flawed update pushed system-wide to sabotage by a disgruntled employee with high-level access to the network, according to Enderle.
Amazon is best known as an online retailer but the Seattle, Washington-based company is also a major provider of cloud-computing services, renting out space on its powerful servers to customers around the world.
Global technology companies are investing billions of dollars in cloud computing, which involves hosting information on the Web and providing it to customers on demand.
"This won't derail the shift to the cloud," Enderle said of Amazon EC2 troubles. "It will just remind people the cloud isn't bulletproof and you need redundancy, either into another cloud service or your own proprietary systems."
Explore further: 'Homebrew' email servers: Genius as well as sneaky?