@nebius/js-sdk - v0.2.3
    Preparing search index...

    Interface InstanceSpec

    interface InstanceSpec {
        "[customJson]"?: () => unknown;
        "[unknownFieldsSymbol]"?: Uint8Array<ArrayBufferLike>;
        $type: "nebius.compute.v1.InstanceSpec";
        bootDisk?: AttachedDiskSpec;
        cloudInitUserData: string;
        filesystems: AttachedFilesystemSpec[];
        gpuCluster?: InstanceGpuClusterSpec;
        hostname: string;
        localDisks?: LocalDisksSpec;
        networkInterfaces: NetworkInterfaceSpec[];
        preemptible?: PreemptibleSpec;
        recoveryPolicy: InstanceRecoveryPolicy;
        reservationPolicy?: ReservationPolicy;
        resources?: ResourcesSpec;
        secondaryDisks: AttachedDiskSpec[];
        serviceAccountId: string;
        stopped: boolean;
        [key: symbol]: Uint8Array<ArrayBufferLike> | (() => unknown) | undefined;
    }

    Indexable

    • [key: symbol]: Uint8Array<ArrayBufferLike> | (() => unknown) | undefined
    Index

    Properties

    "[customJson]"?: () => unknown
    "[unknownFieldsSymbol]"?: Uint8Array<ArrayBufferLike>
    $type: "nebius.compute.v1.InstanceSpec"
    bootDisk?: AttachedDiskSpec

    Specified boot disk attached to the instance.

    cloudInitUserData: string

    Data in cloud-init format for customizing instance initialization. For details, see https://docs.nebius.com/compute/virtual-machines/manage#user-data

    filesystems: AttachedFilesystemSpec[]

    List of Shared Filesystems attached to the instance.

    If you want to interconnect several instances in a GPU cluster via NVIDIA InfiniBand, set the ID of an existing GPU cluster. You can only add the VM to the cluster when creating the VM. For details, see https://docs.nebius.com/compute/clusters/gpu

    hostname: string

    Instance's hostname. Used to generate default DNS record in format <hostname>.<network_id>.compute.internal. or <instance_id>.<network_id>.compute.internal. if hostname is not specified.

    localDisks?: LocalDisksSpec

    Local disks are meaningfully different from regular (remote) disks: they are provided by the underlying host and are tied to a particular VM run. Local disk data is not preserved across Stop-Start initiated via Compute API. Local disks are not provided by default. To get them, explicitly request them via this field. Availability depends on the selected platform, preset and region. Changing this field will result in disks change and content loss, but only after stop and start the instance.

    networkInterfaces: NetworkInterfaceSpec[]

    List of network interfaces attached to the instance.

    preemptible?: PreemptibleSpec

    Include these parameters to create a Preemptible VM and omit them to create a Regular VM For details, see https://docs.nebius.com/compute/virtual-machines/preemptible

    recoveryPolicy: InstanceRecoveryPolicy

    Recovery policy defines how the instance will be treated in case of a failure. Common source of failure is a host failure, but it can be any other failure. Instance undergoing a guest shutdown (poweroff, etc.) will be subject to recovery policy, meaning that it could be restarted and billed accordingly. Stop instance via API or UI to stop it to avoid recovering.

    • If set to RECOVER, instance will be restarted, if possible. It could be restarted on the same host or on another host.
    • If set to FAIL, instance will be stopped and not restarted.
    reservationPolicy?: ReservationPolicy
    resources?: ResourcesSpec

    Specification of compute resources allocated to the instance. For details, see https://docs.nebius.com/compute/virtual-machines/types

    secondaryDisks: AttachedDiskSpec[]

    List of additional data disks attached to the instance beyond the boot disk.

    serviceAccountId: string

    Unique identifier of the service account associated with this instance. For details, see https://docs.nebius.dev/en/iam/about-iam/concepts/service-accounts

    stopped: boolean

    Indicates whether the instance should be stopped.