AWS CloudFront outage knocks websites worldwide offline as 5xx errors…

AWS CloudFront suffered a 3.5-hour outage affecting VPC Origins connectivity, causing 5xx errors across thousands of websites including HuggingFace, the UK National Lottery, and Tailscale. The root cause was a packet processing subsystem failure preventing edge locations from routing requests to customer VPCs. If your infrastructure depends on AWS CloudFront with VPC Origins, you just watched your entire customer-facing availability vanish because of a single routing subsystem bug in one vendor's network. The real cost isn't the three hours—it's that you have no redundancy against the vendor you chose for convenience.