Fortnite and Roblox, two of the biggest games and platforms on the services, were stopped by an AWS outage lasting 15 hours, which was referred to as a severe software defect.

Amazon Web Services (AWS) has unveiled that the mass outage that took place last week was a result of one software bug that brought leading gaming platforms such as Fortnite, Roblox, and PlayStation Network down. Other non-gaming services were affected by the 15-hour disruption as well, and it was one of the biggest disconnects ever measured by Downdetector.

Amazon engineers state that the problem started in the DNS Enactor system of DynamoDB that handles domain lookups. A series of cascading software breakages caused an intermittent DNS status, and an investigation into this state had to be fixed manually.

The power outage underscored the intensive use of the cloud-based infrastructure by the gaming industry, with millions of gamers reporting losing their Internet connection in such games as Rocket League or Pokemon Go. Ookla network intelligence registered the incident as one of the most momentous in the history of AWS.

Even though Amazon has maintained stability since, the incident highlights the instability of internet-based ecosystems that rely on cloud-based games. It is predicted by industry members that the upheaval could lead the developer to consider multi-cloud models that prevent future failures.