Amazon's servers couldn't handle the flood of traffic from shoppers during Prime Day – and the company cut off all international traffic and launched a backup landing page in an effort to fix the problem, according to internal documents obtained by CNBC. Amazon declined to comment to CNBC and did not immediately respond to request for comment from Business Insider. At exactly at 12pm. PDT on Monday, right after Amazon's annual doorbuster sales event began, the company's website was riddled with glitches and outages. A source told CNBC that the scene inside Amazon was “chaotic” and about 300 people got on an emergency conference call to try and deal with the issues. According to outside experts who reviewed the documents for CNBC, Amazon's auto-scaling technology may have broken down under the burden, which would have limited the number of servers available to respond to the boom in traffic. Auto-scaling automatically increases or decreases cloud server capacity based on changes in traffic. During the first minutes the site was experiencing glitches, Amazon created a “scaled-back” front page to relieve pressure on servers, CNBC reported. At the time, users who clicked on buttons on Amazon's main landing page weren't taken anywhere. The company also cut off all international traffic in an effort to reduce server load. The reduced server capacity allegedly started to effect another tool called Sable, which coordinates several other internal infrastructure services at Amazon. At about 1pm PDT, says the report, Sable sounded a “red” emergency alert, which could have been related to problems at other parts of the site, including the Alexa voice service and Prime Video. At the time, the company acknowledged the technical hiccups in a short statement, which said “some customers are having difficulty shopping, and we're working to resolve this issue quickly.” But despite the embarrassing glitches, Amazon reported that Prime Day saw a record number of sales, becoming the biggest shopping day in the company's history.