On October 11, a failure occurred in Yandex Cloud that affected the Layero infrastructure. The control panel, deployments and sites on Layero have become unavailable.
On October 18, the incident was closed. The control panel, builds, deployments and sites are working. Recovery history - below.
By Yandex Cloud message, as a result of a UAV attack, the infrastructure of the data center was damaged. The incident affected the services on which Layero was running.
The operation of the panel, new deployments, sites, payments and support has been restored. Current component availability is posted on status page.
Management services, server applications and object storage have been moved to Selectel. Public website addresses are saved; There is no need to change DNS due to this transfer.
For new users, the free period is again 14 days. Free period extensions and payment transfers already granted will remain in place. The term of your period and the date of the next payment can be checked in your account dashboard.
The old Shared database data has not yet been restored; their safety and restoration period have not yet been confirmed. Closing an incident relates to the operation of the service. If you have questions about your database data, please contact support.
Latest updates at top. Time is indicated in Moscow.
Emergency API restrictions have been lifted, background tasks have been resumed. The 14-day free period has been restored for new users. Previously issued extensions are retained. The issue of restoring old Shared data remains open.
The panel and documentation have been transferred to Selectel S3 storage. We checked HTML, JavaScript, styles, deep panel paths and 404 responses. The transfer of custom deployment artifacts continues.
Linking a payment method and canceling auto-renewal are available again.
Reception of requests, delivery of responses to the office and mail notifications have been restored. We checked both directions of exchange: from the account to support and back.
We checked the restoration of the platform database and copies of the Shared databases of the new circuit in a separate environment. The check applies to new infrastructure and does not confirm the return of old Shared data from the emergency data center.
Opened a account dashboard for all users. Access to dedicated databases has been restored; working with old Shared data remains a separate step.
We checked login via Yandex and email, import of repositories, processing of Git-webhooks and auto-deployment. We checked the basic API operations through the panel, CLI and MCP.
Prepared the Selectel registry for application images. We checked the loading of images using the checksum and the launch of containers. Previously published images were migrated separately from new builds.
The end-to-end test tested the creation of a deployment, queue processing, build and launch of a fullstack application. The frontend and backend API responded to the public HTTPS address.
Directed incoming web traffic to new Selectel nodes. After loading errors, we rolled back the switch, adjusted the network parameters and re-checked HTML, JavaScript, styles and HTTPS. Public addresses have been preserved.
The test fullstack application was built and launched on the new infrastructure. We checked the frontend, server API and separately SSR.
Connected the existing subassembly to the rebuilt platform. It began receiving tasks from the queue of the new control loop.
Restored the platform base and prepared the main services in Selectel. Recording operations and background tasks were opened in stages after checks to prevent the old and new circuits from running at the same time.
Restored distribution of static sites and loading of the control panel. The availability of the panel did not mean that the API, new deployments and server applications were fully operational.
External checks showed timeouts for the panel, API and user sites. 200 responses from individual S3 sources and balancer checks did not confirm platform availability for users. Disaster recovery has begun.
The start time is indicated by Yandex Cloud status. The provider reported that the zones were unavailable. The incident affected the infrastructure on which Layero's management services, websites and user databases worked.