Knowledge Base

Introduction to Jobs

Introduction to Jobs

Jobs are tasks within the Narrative platform. Every operation you perform—like running a forecast or executing a query—leads to the creation of a job. Jobs allow for tracking of these operations and their status.


What Are Jobs?

Jobs represent discrete units of work performed by the Narrative platform. Each job has:

  • Unique identifier: Every job can be tracked individually
  • Type: Indicates what kind of work the job performs
  • State: Shows the current status of the job
  • Input parameters: Data and settings that control job execution
  • Results: Outputs produced when the job completes successfully

Job Characteristics

Trackable Operations

Jobs provide visibility into platform operations that might otherwise be invisible. Whether you're creating a dataset or running a forecast, the corresponding job lets you monitor progress and troubleshoot issues.

State-Based Progression

All jobs progress through well-defined states, making it predictable to understand where any operation stands in its lifecycle.

Parameterized Execution

Jobs accept specific input parameters that control how they execute, allowing for customized behavior based on your needs.


Common Job Operations

Based on the types available in the platform, jobs handle several categories of work:

Data Processing

  • Creating and refreshing materialized views
  • Running forecasts to estimate query results
  • Calculating costs for data processing operations

Dataset Management

  • Generating data samples for preview
  • Calculating column statistics for data profiling
  • Cleaning up data resources
  • Delivering data to external systems

Model Operations

  • Training machine learning models
  • Deploying models to production environments

Job Lifecycle Overview

Every job follows a consistent lifecycle:

  1. Creation: Job is submitted with specific parameters
  2. Queuing: Job waits for available resources
  3. Execution: Job runs and performs its work
  4. Completion: Job finishes with success or failure
  5. Cleanup: Resources are released and results stored

Job States

Jobs progress through these states:

StateMeaning
pendingJob is queued, waiting to start
runningJob is actively executing
completedJob finished successfully
failedJob encountered an error
cancelledJob was cancelled before completion
pending_cancellationJob is being cancelled

Why Jobs Matter

Understanding jobs helps you:

  • Monitor Operations: Track the progress of long-running operations
  • Troubleshoot Issues: Identify and resolve problems when they occur
  • Optimize Performance: Understand execution patterns and resource usage
  • Plan Resources: Anticipate when operations will complete

Next Steps

Now that you understand what jobs are, explore these topics:

Job Types Reference

Learn about the specific types of work that jobs can perform.

< Back
Rosetta

Hi! I’m Rosetta, your big data assistant. Ask me anything! If you want to talk to one of our wonderful human team members, let me know! I can schedule a call for you.