On July 16, 2026, the HEFLO platform experienced a period of instability during which some users were unable to access the application. The disruption was caused by a failure in a component of our cloud provider's content delivery infrastructure (AWS CloudFront VPC Origins). We restored service the same day by routing traffic through an alternate path, and have since implemented a redundant architecture to prevent a recurrence.
During the incident window, affected users encountered errors (including HTTP 504 Gateway Timeout) when accessing the platform. The issue affected multiple regions simultaneously. Service was fully restored on the same day.
The disruption originated from a failure in AWS CloudFront VPC Origins, the mechanism that allowed CloudFront to reach our application over a private, non-public connection. This is a provider-side infrastructure component; the failure was not caused by a change on our side. Because the affected component sits in front of all applications, the impact was broad rather than isolated to a single service.
Our team established an alternate delivery path that bypassed the affected component, restoring access the same day. We then validated end-to-end functionality across regions before closing the incident.
Over the following days we implemented a redundant delivery architecture. Traffic now has an automatic failover path: if the primary route becomes unavailable, requests are automatically redirected to a secondary path with no manual intervention required. This significantly reduces the time to recover from any similar provider-side disruption in the future.
We appreciate your patience and understanding during this incident.