Job Settings

This topic presents the setting options available for your job.

You can access the job settings from the The "Overview" page icon is a square divided into several other squares. Overview page of the job.

Job settings from the "Overview" page.

It includes the following non-versioned and editable features:

1. Name

Names are mandatory. They can be up to 255 characters long and must be unique within a project.

2. Alias

The job alias is unique to each job in a project. It allows you to reference a job within another job and can be used to pass information between jobs during pipeline execution. However, to make this work, the settings env vars Variables setting must be enabled.

3. Description

Descriptions are optional and not restricted. However, it is a good practice to add them. Keep them short and to the point.

You can also use ChatGPT to generate your job description by clicking Generate with ChatGPT (a). For more information, see Generating a Job Description With ChatGPT.

This feature is only available for Spark with a Python context, Bash, R, Sqoop, and Python job technologies. It works with .py, .sh, .r, and .bash files and command lines, but it does not work with archive files.

4. Scheduled Run Icon for runtime types is a calendar.

There are two types of execution:

  1. The manual run, which requires you to click Run Run to start the job.

  2. The scheduled run, which launches the job according to the schedule you choose. It has three schedule modes: Simple, Shortcut, and Expert.

    A scheduled job can always be run manually.
    • Simple Mode

    • Shortcut Mode

    • Expert Mode

    In Simple mode, you can easily specify variables through the user interface. There are many options.

    Screenshot of the settings for the scheduled run type in simple mode

    In Shortcut mode, you can choose the recurrence of your run on an hourly, daily, weekly, monthly, or annual basis. All other settings are automatic.

    Screenshot of the settings for the scheduled run type in shortcut mode

    In Expert mode, you can specify variables using the Cron format. The Cron time string consists of five values separated by spaces: [minute] [hour] [day of the month] [month] [day of the week]. They are based on the following information:

    Table 1. Cron format
    Descriptor Acceptable values

    Minute

    0 to 59, or *

    Hour

    0 to 23, or *

    Day of the month

    1 to 31, or *

    Month

    1 to 12, or *

    Day of the week

    0 to 7 (0 and 7 both represent Sunday), or *

    The Cron time string must contain entries for each character attribute. If you want to set a value using only minutes, you must have asterisk (*) characters for the other four attributes that you are not configuring.

    Screenshot of the settings for the scheduled run type in expert mode

    Regardless of the Scheduled run type selected, you can specify the time zone for your jobs or pipelines. When specifying its time and day, you should know that:

    • The default value in the time zone selection is your browser’s time zone.

    • The next schedule preview will be displayed in the time zone of your browser.

    • Daylight Saving Time (DST) is automatically adjusted.

    • Runs scheduled for the 29th, 30th, or 31st day of the month will not run in months with less than 29, 30, or 31 days.

      Example 1. French time settings

      UTC time is two hours behind French time, so if you want your job to run at midnight French time, set it to two hours later, that is, at 2 a.m.

    Once you have finished scheduling your run, you will see the summary of your choice written below and the time of the next run.

By default, instances can run at the same time. To prevent this, select the Forbid overlapping scheduled instances option. This means that if your job is still running when the next scheduled run is due, it will be skipped.

For Spark jobs only.

To prevent collisions due to concurrent executions of one Spark job, the orchestration cannot be scheduled under one hour.

job settings spark

5. Alerts The "Alerts" icon is a bell.

Alerts are optional and can be set to receive an email when the status of your job changes. They can be sent to multiple email addresses to notify you of the following status changes:

Status Description

queued requested Requested

The job’s run has been requested and is being executed.

queued requested Queued

The job is waiting for the necessary resources to be executed.

spinner Running

The job is up and running.

fail Failed

The job has crashed.

A failed job can go into an out of memory Out Of Memory (OOM) state, which is an extension of the Failed state. The OOM state can be due to a lack of memory (RAM).

stop Stopping

The job is stopping.

stop Stopped

The job has stopped running.

success Succeeded

The job has been successfully executed.

unknown Unknown

The job no longer runs because an error has occurred.

6. Resources Icon for CPU and RAM resources.

For embedded jobs only.

CPU, RAM, and GPU resources are optional, but are recommended. They can be specified for optimal execution.

The consumption of your job can be managed by guaranteed resources, that is, the minimum amount of resource requested, and limited resources, that is, the maximum amount of resource that can be consumed.

When you create your job, CPU and RAM resource management is enabled by default with predefined values. In other words, based on the default values defined at technology level, Saagie automatically assigns resource requests and limits to your job. These values can be adjusted to suit your needs.

Default values are already set by Saagie at the technology level. These values are mandatory. They also exist at the technology context level, where they can override the values defined at the technology level. You can configure them when you create your job, or by modifying the Resources Icon for CPU and RAM resources. setting of the job.

If the resource capacity is not defined, the defaults that are defined at the technology level will be assigned at the technology context level.

Automatic adjustments can be made to avoid inconsistent configurations. If you try to set a guaranteed value that is greater than the limit value, a note appears to inform you that the guaranteed value has been adjusted. If you try to set a limit value smaller than the guaranteed value, a note appears to inform you that the limit value has been adjusted (a).

RAM limit adjustment message.

In addition, if the guaranteed value and the limit value are not optimal, a message is displayed with the appropriate values for an optimal configuration (b).

CPU recommendation message.

To enable GPU management on your job, select the Run on GPU option. You do not need to configure it, as a default value is set for both limited and guaranteed resources. This value cannot be changed.

GPU option selected.
When you change the CPU, RAM, and GPU resources, your job is automatically restarted.