AWS USEast1 DNS failure disrupts DynamoDB, triggers wide outages

Amazon Web Services suffered a major outage early on Monday, October 20, centered in its USEast1 region in Northern Virginia, that left thousands of companies and millions of users unable to access popular apps, games and online services worldwide. The disruption began in the predawn hours and, according to AWS, stemmed from a Domain Name System failure that prevented clients from reaching DynamoDB, a core AWS database service, triggering cascading errors across many dependent systems.

AWS first reported increased error rates and latency for multiple services in USEast1 and later identified problems with DNS resolution for the DynamoDB API endpoint. When DNS lookups failed, applications and AWS internal components were unable to locate DynamoDB endpoints, and other services that rely on those connections, including EC2 and Lambda, experienced impaired operation. Engineers manually corrected an incorrect DNS state, disabled parts of the DNS automation while repairs were made, and applied throttling to stabilize recovery and process large backlogs of queued requests.

The outage affected a broad range of consumer and enterprise platforms. Users reported interruptions to social and...