Skip to main content

Preemptible cloud servers

A preemptible cloud server is a cloud server that runs for no more than 24 hours after creation and can be stopped by Selectel at any time, for example, if the virtual host does not have enough resources for other cloud servers.

During a system preemption, the cloud server is not deleted — it is stopped and switches to the EXPIRED status. After preemption, the server can be restored. All data is preserved on a server with a network boot drive. Data is deleted from local drives of preemptible servers.

Preemptible servers support all features available for standard cloud servers, while their cost is lower. Preemptible cloud servers are available in pools in Russia.

You can make a cloud server preemptible when creating a server or change the type of an existing server. A server of any configuration can be made preemptible.

Use cases

Suitable for fault-tolerant systems where multiple servers are used and when some of them fail, the load is redistributed to other servers:

  • for parallel batch data processing;
  • CI/CD testing;
  • Hadoop and Kubernetes projects;
  • scaling fault-tolerant web services during peak load times;
  • any fault-tolerant projects with variable load.

Limits and constraints

We do not guarantee the same availability level as for standard cloud servers — preemptible servers are not covered by the Cloud Platform SLA .

Cost

The cost of a preemptible server is, on average, 70% lower than the cost of a regular cloud server with the same configuration.

While running, preemptible cloud servers are billed according to the Cloud Platform billing model.

After interruption:

  • vCPU, RAM, GPU, and local disks are no longer charged starting from the next hour after the stop;
  • public IP addresses, public subnets, and network volumes continue to be charged.