How a DNS Failure Took Down the Internet
AWS first reports operational issue in US-EAST-1
Over 14,000 reports for Amazon.com
DNS resolution failure identified
Notable signs of recovery reported
Global services resume operation
"Perplexity is down right now. The root cause is an AWS issue. We're working on resolving it."
Model training disrupted
Inference unavailable
Query processing failed
Database queries failed
Backend functions unavailable
Microservices failed
Notifications stopped
User data unavailable
Content unavailable
Database queries failed
Event processing halted
Slow loading
A DNS resolution failure in the DynamoDB API endpoint triggered a domino effect across 15+ AWS services and 32+ major applications worldwide.
Located in Northern Virginia, US-EAST-1 is AWS's first region with legacy architecture dating back to early AWS days, making it more complex and harder to maintain.
AWS's microservices architecture means services depend on each other. When DynamoDB failed, services like Lambda, EC2, and IAM that depend on it also degraded.
Despite being a regional issue, the outage affected users globally because many services use US-EAST-1 as their primary region or for global features.