Skip to main content
Nebius AI Cloud performs maintenance of Managed MLflow clusters. This page explains how to get information about the maintenance schedule and reschedule maintenance operations.

​
Prerequisites

  1. Install and configure the Nebius AI Cloud CLI.
  2. Check that your project ID is saved in the Nebius AI Cloud CLI profile configuration: 
    cat ~/.nebius/config.yaml
  3. Install jq to extract IDs and tokens from the JSON data returned by the Nebius AI Cloud CLI: 
    sudo apt-get install jq

​
How to get a list of all maintenance operations

Run the following command to get a list of all past and upcoming maintenance operations: 
nebius maintenance v1alpha1 maintenance list
This command returns a list of all maintenance operations. You can use this information to make sure that you are not relying on a resource that is unavailable due to maintenance. Each item contains:
  • metadata.id: The ID of the maintenance operation. You can use this ID to get the details of an operation or edit it.
  • metadata.parent_id: The ID of the project that contains the resource affected by the maintenance.
  • metadata.created_at: The time that the maintenance operation was created.
  • spec.scheduled_at: The time that maintenance is scheduled for, in the ISO 8601 format.
  • status.affected_resources: The identifiers of resources affected by the maintenance and the ID of the project they belong to (parent_id). If the maintenance event affects your resources in another service, they are also listed here. You can filter the output to see only the affected resources for .
  • status.state: The state of the maintenance, for example, STATE_FINISHED, STATE_SCHEDULED or STATE_CANCELLED.
  • status.reschedulable: If set to true, and status.state is set to STATE_SCHEDULED, you can reschedule the maintenance.

​
Filtering the output

To filter the list output by MLflow clusters, run the following command: 
nebius maintenance v1alpha1 maintenance list \
  --format json \
| jq '
  .items[]
  | select(any(.status.affected_resources[]?; .id | startswith("mlflow-")))
  | .status.affected_resources |= map(select(.id | startswith("mlflow-")))
'
This command returns only maintenance operations that affect MLflow clusters, and removes resources belonging to other Nebius AI Cloud services from the affected_resources array of each list item.

​
How to get the details of a maintenance operation

To get details of a maintenance operation, run the following command: 
nebius maintenance v1alpha1 maintenance get --id <maintenance_operation_ID>
This command returns the details of a maintenance operation. The details are the same as for the list maintenance operations command. 
metadata:
    id: maintenance_***
    parent_id: project-***
    created_at: "2025-09-22T14:40:59.845329Z"
spec:
    scheduled_at: "2025-10-01T14:28:20Z"
status:
    affected_resources:
      - id: mlflow-***
        parent_id: project-***
    started_at: "2025-10-01T14:28:21.162373Z"
    finished_at: "2025-10-01T14:34:38.363547Z"
    state: STATE_FINISHED
    reschedulable: true

​
How to reschedule a maintenance operation

You can reschedule a maintenance operation if it meets the following conditions:
  • status.state is set to STATE_SCHEDULED.
  • reschedulable is set to true.
To reschedule maintenance:
  1. Run the following command: 
    nebius maintenance v1alpha1 maintenance edit --id <maintenance_operation_ID>
  2. In the editor that opens, change the value of spec.scheduled_at to your desired time in the ISO 8601 format. Keep in mind that you can only specify a date and time that is in the future.
  3. When prompted, confirm the update.
After you reschedule a maintenance operation, use the get command to check the updated details.