What happened
In a press release dated 23 October 2026, Yandex said Yandex Cloud engineers had restored access to 90.47% of the data stored on the platform after attacks on Yandex data centers. The company framed this as an access-restoration figure; it did not say that every affected service or customer workload was fully back to normal.
Some data and services associated with them were moved to a new location that began operating on 23 October. Yandex said access to this migrated material would be opened in stages over the following days. Customers would learn that access had become available through Yandex Cloud mailing-list notifications. The release does not provide a schedule by customer, dataset, or service, so the stated staged rollout is the operational detail available to users rather than a firm individual recovery deadline.
Yandex said the platform’s main capabilities were already available. Customers could create virtual machines and databases, manage Kubernetes clusters, use business analytics and cybersecurity services, and use Yandex AI Studio tools for AI work. The statement distinguishes those main capabilities from other services still being brought back: the team was continuing to establish the operation of additional services while restoring the remaining data.
The source does not list the data-center locations or quantify how much data remained inaccessible in each one. A separate supplied event summary says the disruptions began with the shutdown of the Sasovo data center and later the Vladimir site, affecting zones ru-central1-b and ru-central1-a, respectively. Those details provide context for the incident, but the 23 October release itself does not describe the outage sequence or explain the technical recovery method. It also gives no price, customer count, or service-level impact breakdown.
What the archive adds
This is a new recovery fact in the ongoing Yandex Cloud incident, not a restatement of the initial disruption. Yandex Cloud arranged accelerated placement of data and services for affected customers described a cross-provider recovery path; this update gives a platform-level data-access figure. A separate Yandex update on Yandex Disk access concerned a different service and should not be conflated with Cloud’s 90.47% figure.
Trend signal
The update supports agent security moves to runtime boundaries only indirectly: cloud reliability is a prerequisite for systems that execute and store agent work, but this release offers no evidence about agent runtime security itself. The more direct signal is operational resilience: a provider’s recovery path, service availability, and data-access restoration are distinct dimensions that customers need to track separately.
Why now
The release is dated 23 October, the day the new location began operating. It records the next recovery milestone after the data-center disruptions and explains that access to migrated data would be rolled out over the following days.
Connections to other events
The archive’s earlier account of Yandex’s response to the data-center damage described accelerated placement of affected customers’ data and services with another provider. The present update adds a different measure: Yandex’s own statement that access had returned to 90.47% of Yandex Cloud data, with remaining data and services still under restoration. These are related parts of one incident response, not interchangeable recovery metrics.
Why it matters
For a team using cloud infrastructure, “the platform is available” and “our data and services are accessible” are not the same claim. Yandex’s update shows why recovery communications should be read at the level of the specific resource: the company says major platform capabilities are back, while a portion of data and other services remain in recovery and access to migrated material is being phased in.
Teams should use provider status updates as prompts to verify their own workloads rather than assuming a published aggregate percentage applies to them. Check whether critical data is accessible, whether dependent services operate, and whether the recovery path requires customer action. The release offers a useful milestone, but does not establish customer-by-customer recovery or provide enough detail to assess the resilience of any particular workload.
Limits and uncertainty
The 90.47% figure is Yandex’s own aggregate statement; the release supplies no underlying methodology, denominator, independent verification, or breakdown by customer, region, service, or data type. It does not define what “access restored” means for partially available datasets or dependent workloads. Nor does it give a date for completion of the remaining recovery. The announcement confirms that a new location began operating and that access would open in stages, but it does not explain the architecture, replication strategy, or cause of the disruptions in this release.
Reader takeaway
Yandex’s 90.47% figure is a meaningful recovery milestone, not proof that every customer’s data or services are back. Treat platform-level recovery announcements as a cue to check your own workloads, data access, and dependencies—and keep tracking the remaining service restoration.
Open question
When will Yandex Cloud finish restoring the remaining data and services, and how will it report recovery at the level customers can verify?