Major AWS Outage in US-East-1 Disrupts Global Internet Services; DNS Issue Cited as Root Cause

October 20, 2025 – A widespread disruption originating in Amazon Web Services’ (AWS) critically important US-East-1 region has caused major outages and connectivity issues for countless websites and applications around the world. The incident, which highlights the global dependence on AWS cloud infrastructure, impacted services across retail, finance, social media, and gaming sectors.

AWS confirmed that the root cause of the failure appears to be related to the DNS (Domain Name System) resolution of the DynamoDB API endpoint within the US-East-1 data center in Northern Virginia. DynamoDB, a key NoSQL database service, underpins numerous other AWS functions, leading to a cascading service failure that affected the entire region and global services dependent on it.

Widespread Impact Across Key Services

The disruption led to inaccessibility and elevated error rates for a diverse range of high-profile platforms. Among the services reported to be affected were:

• Amazon Ecosystem: Amazon.com, Prime Video, and the voice assistant Alexa.
• Social & Communication: Snapchat, Signal, and the learning app Duolingo.
• Financial Services: Cryptocurrency exchange Coinbase, the trading app Robinhood, and the payment platform Venmo.
• Gaming & Media: Roblox and Fortnite.
• Business & AI: Perplexity AI, and the design platform Canva.

The outage demonstrated how failures in a single, vital AWS region can quickly generate massive ripple effects, paralyzing services that operate far outside the affected physical data center.

Recovery Efforts Underway

AWS engineers were immediately engaged to mitigate the issue. According to the latest updates from the AWS Health Dashboard, initial mitigations have been successfully applied, and the company is observing “significant signs of recovery” across most affected AWS services.

While many platforms are showing signs of restored functionality, AWS noted that services are working through a “backlog of queued requests,” meaning users may still experience intermittent latency and minor issues as full operational stability is restored. The company continues to monitor the situation and provide updates as the recovery process moves toward full resolution.