Yandex Cloud Suffers Third Data Center Attack

This is a brutal reminder that availability zones only protect against independent failures; coordinated physical attacks can collapse the assumptions underneath them. For AI and cloud developers, the incident turns disaster recovery from an architecture diagram into an immediate operational test.

• –ru-central1-a is fully unavailable, while ru-central1-b also remains offline from an earlier attack, leaving Yandex Cloud operating with sharply reduced redundancy.
• –Compute remains available in ru-central1-d and ru-central1-e, but managed databases, Kubernetes clusters, and Object Storage face degraded or read-only operation.
• –Yandex recommends multi-zone deployment, yet this incident shows why critical workloads also need off-provider backups and recovery paths.
• –AI developers relying on Yandex Cloud should test restore procedures, quota constraints, cross-provider failover, and dependencies on regional control-plane services.