Skip to main content

Celonis Product Documentation

Scheduling machine learning notebooks

Machine learning notebooks can be executed on a recurring schedule within the Celonis Platform, allowing you to control when they execute, how long the timeout is, and what the maximum number of retries should be.

Once notebooks are schedule and running, you have a number of management options: Managing existing machine learning notebooks schedules

To schedule your machine learning notebooks:

Important

For your notebook to be successfully scheduled, your workbench must contain a valid execution file.

  1. Click Data - Machine Learning.

    Data_-_Machine_learning.png
  2. Click Scheduling and then click New Schedule.

    new_schedule.png
  3. Configure your schedule, including name, which workbench and execution file to use, and your frequency options.

    Your options here are as follows:

    • Time

      • Hourly (Full hour, quarter past, half past, quarter to)

      • Every few hours (with customizable number of hours, plus the above hourly options from there)

      • Daily (with customizable time)

      • Weekly (select 1 or multiple days and the time the notebook should run)

      • Monthly (select calendar day and the time the notebook should run)

      • Custom cron (freely define a scheduling plan by using space-separated values, with 1 minute being the minimum period between runs)

    • Configuration / Frequency options

      • Execution timeout (select from minutes or hours)

      • Maximum retries (select the number of retires before the notebook stops running)

      • Receive emails for failed executions (with an email summary sent to the email address you are accessing your Celonis Platform team with)

    new_schedule_creation.png
  4. Click Save.

    Your schedule is configured and set to disabled, meaning that it will not currently run to your configured frequency. To enable this, click the Options button and then click Enable.

Managing existing machine learning notebooks schedules

Once scheduled, you can manage your machine learning notebook schedules by clicking the options button:

manage_existing_ml_notebook_schedules.png

Your options here include:

  • Run: Manually run or execute you notebook once, regardless of your configured frequency and whether your schedule has been enabled.

  • Edit: Update all details of your schedule and save any changes you make.

  • Delete: Disable and delete this schedule from your list, with no reversal or restoration possible.

  • Permissions: Configure notebook level permissions only.

  • Enable / Disable: Once enabled, your schedule will run to your configured frequency.

  • Subscribe / Unsubscribe: Receive email notifications whenever your schedule runs.

    As an example, this is an email received when a run fails:

    failure_scheduling.png