OVH SAS

OVHcloud's October 2021 outage began with planned work intended to strengthen anti-DDoS protection, but the public record describes a routing-control failure rather than a successful attack. A command affecting redistribution from BGP into OSPF caused the full Internet routing table to enter the internal routing system, exhausted control-plane resources, broke IPv4 reachability, defeated the attempted rollback, and forced engineers to isolate and power off a router before the network could converge. The incident is a test of whether production change controls govern the routes that actually run, not merely the procedure that was approved.

The dependency surface is workload hosting and network reachability. Customers may use OVHcloud for dedicated servers, public cloud, private cloud, or domain/hosting services. If data center availability, backbone routing, product reliability or group investments change, customers may face migration and continuity consequences.

A 17-minute OVHcloud disruption in October 2024 shows why routing resilience cannot be assigned to only one side of a peering session. The event joined an erroneous outbound announcement to a missing inbound limit, exposing a control interface where responsibility must be both separate and shared.