Facebook blames server slipup for longest outage in its history

Facebook's daylong outage has finally ended. 

The troubles first hit the social network -- which has more than 2 billion users around the world -- around 9 a.m. PT Wednesday, according to outage tracking site Downdetector. At this point on Thursday, more than 24 hours later, users have by and large regained full access.

On Thursday, Facebook blamed the problem on "a server configuration change," saying that it had resolved the issues and that "our systems are recovering."

Instagram was also down for many hours, starting at around the same time. However, Instagram's official Twitter account announced the platform was back in action at 9:41 p.m. PT time on Wednesday.

Users throughout the day Wednesday posted on  Twitter  that they were seeing a message saying Facebook was "down for maintenance." Other users trying to post status updates on Facebook, including CNET staff, got error messages saying "something went wrong." 

The outage mainly affected users in the US and Europe but also hit parts of South America, Australia and Asia, according to a map by Downdetector. 

Users of Facebook-owned WhatsApp also reported they were having trouble sending photos on the app. Virtual reality company Oculus, also owned by Facebook, reported that its users, too, were having trouble accessing and using the platform. 

Watch this: Where were you when Facebook went down?

01:42

While Facebook has experienced lengthy outages in the past, downtime like this is unusual among the world's largest tech companies.

Downdetector received about 7.5 million reports from Facebook, WhatsApp and Instagram users during the outage, making it the largest outage the tracking site has seen since it was launched in 2012. A YouTube outage in October came in second, with 2.7 million reports.

"As people become more reliant on digital services like Facebook, Instagram and YouTube, we are witnessing an ever growing number for problems during outages," said Tom Sanders, co-founder of Downdetector, in a statement. "The tolerance for downtime decreases and people are increasingly expecting services to operate flawlessly 365 days per year."

Facebook, which makes money from selling ads, was having issues with its ad products as well. When asked if Facebook would be refunding advertisers for the outage, a company spokesman said that Facebook isn't sharing any details about refunds at this time.

Facebook had said in a tweet Wednesday that the issue didn't arise from a type of cyberattack known as a distributed denial-of-service (DDoS) attack, in which hackers try to crash a site by flooding it with more traffic than its servers can handle. 

In November, Facebook went down in a mishap caused by a test the company was running. That outage lasted about 40 minutes.

That same month, Facebook and Instagram were down for hours because of what Facebook said was a server configuration. This is, however, the longest outage in the company's history. The previous record-holding outage occurred in 2008, when Facebook had just 125 million users.

Originally published March 13 at 10:22 a.m. PT.
Updates, 10:46 a.m.: Adds comment from Facebook; 12:18 p.m.: Includes tweets from Facebook; 1:23 p.m.: Adds background on what areas were affected, and info on WhatsApp; 5:06 p.m.: Includes the length of the outage and more info on WhatsApp; 11:19 p.m.: Adds Instagram tweet and length of Facebook's outage; March 14 at 10:11 a.m. PT: Adds Facebook statement and notes that Facebook has recovered. March 14 at 12:51 p.m. PT: Adds remarks and data from Downdetector and comment from Facebook.