AWS CloudFront Outage Knocks Sites Offline as VPC Origins Configuration Failure Cascades Across Regions
A CloudFront routing-configuration failure tied to VPC Origins triggered widespread 5xx errors on July 16, hitting Hugging Face and the UK National Lottery before AWS restored service.
Editor's Note ·
- Correction:
- The Analysis section quotes APIContext CEO Mayur Upadhyaya as saying, "When there's an outage, it impacts a greater number of users." This phrasing is not a verbatim quotation. The cited SDxCentral article states that Upadhyaya "noted the incident itself wasn't the story, but the 'concerning trend' of provider consolidation, meaning that when there's an outage, it impacts a greater number of users" — only "concerning trend" is presented in the source as his exact words; the rest is the reporter's paraphrase, not a direct quote.
Overview
Amazon Web Services suffered a multi-hour CloudFront outage on July 16 after a routing-configuration failure tied to the content delivery network’s VPC Origins feature began throwing 5xx errors, knocking websites and online services offline across regions, according to The Register. VPC Origins is described by The Register as a relatively new CloudFront feature that lets customers serve applications running behind private load balancers without exposing their back-end infrastructure to the public internet, and AWS said customers using other origin types were unaffected.
What We Know
AWS’s status page put the start of the disruption at 0145 PDT (0945 UTC), according to The Register, which first reported the incident. SDxCentral separately reported that increased “5xx” errors occurred between 12:45 a.m. and 4:18 a.m. PDT for CloudFront VPC Origins users, a window of roughly three and a half hours.
AWS’s first public status update read: “We are experiencing increased 5xx errors for CloudFront customers utilizing VPC Origins connectivity. Our engineers are engaged and are actively working to mitigate impact,” according to The Register. At 03:18 PDT, AWS added that “we believe the root cause is related to a packet processing subsystem responsible for routing requests from CloudFront’s edge locations to resources within customer VPCs,” per the same report. Users hitting affected sites saw an error reading, “We can’t connect to the server for this app or website at this time. There might be too much traffic or a configuration error. Try again later, or contact the app or website,” The Register reported.
AWS’s final update, as quoted by The Register, read: “an internal constraint on the fleet that manages connections to private VPC origins. When this constraint was reached, the system responsible for distributing routing configuration to our network processors failed to load the updated configuration data correctly, affecting routing of VPC Origin connections.” SDxCentral reported that AWS said the issue was resolved and that customers who had temporarily switched origin types could safely revert.
Among those affected, Hugging Face acknowledged its service was unavailable “from most regions in the world” due to the AWS outage, The Register reported. The UK’s National Lottery posted on X that players were unable to access its website and mobile app, describing it as a wider AWS outage, according to the same report. Players of Fallout 76 also reported issues on Reddit regarding inaccessibility, The Register noted.
During the outage, AWS recommended that customers who did not strictly need VPC Origins temporarily switch to a different origin type while engineers worked on a fix, The Register reported.
What We Don’t Know
AWS has not disclosed what specifically caused the fleet-management constraint to be reached in the first place, nor has it published a detailed public post-event summary of the incident. The two contemporaneous reports also differ on the outage’s precise start time by about an hour — The Register cites AWS’s own 0145 PDT figure, while SDxCentral cites a 12:45 a.m. PDT start — a discrepancy neither outlet has since reconciled.
Analysis
The incident is the latest reminder of how a single configuration fault inside a hyperscaler’s edge network can ripple across unrelated customers with no direct relationship to one another, from an AI model-hosting platform to a national lottery operator to an online video game. SDxCentral quoted APIContext CEO Mayur Upadhyaya on the broader dynamic: “When there’s an outage, it impacts a greater number of users,” a nod to how concentration in a handful of cloud providers amplifies the blast radius of any individual failure.