• Documentation
    • Search documentation
    karat

    +

    K

    API Reference ↗
    Ontology buildingOverview

    Scenarios [Beta]

    Beta

    Ontology scenarios are in the beta phase of development and may not be available on your enrollment. Functionality may change during active development. Contact Palantir Support to request access to scenarios.

    With Ontology scenarios, you can create and compare "what-if" analyses using objects and the ontological concept of a model. A scenario is essentially a sandbox to apply edits on top of the data in your Ontology, generated by applying one or more actions.

    Common use cases

    We recommend using Ontology scenarios for the following workflows:

    • What-if analysis: Simulate changes to your data or system by creating an isolated fork of your Ontology, without affecting live data.
    • Business logic simulation: Apply actions and functions to simulate business processes or evaluate the impact of operational changes before committing them.
    • Agent evaluation and forecasting: Test agents against different input parameters to compare outcomes and refine predictions.
    • Comparative analysis: Create multiple scenarios side by side to identify the optimal decision or understand the trade-offs between di

    Scenarios are not data versioning tools. They cannot provide a snapshot of your Ontology at a historical point in time and should not be used for that purpose.

    Time to live

    To manage storage costs, a scenario has a default lifespan of 30 days. After 30 days, the scenario is automatically deleted along with all edits applied to that scenario and the backing persisted scenario object.

    Auto-rebasing

    A scenario is auto-rebased every 10 minutes on top of their main or global branch. The rebase feature allows you to incorporate the latest changes from the base Ontology into your active scenario, keeping it in sync with the latest state of your data.

    Global Branching vs. Ontology scenarios

    Global branching is a tool used by builders to develop and test end-to-end workflows in the Palantir platform. These workflows would otherwise be too disruptive to run against a live production environment. An Ontology scenario provides a sandbox for users (humans and agents) of these workflows to apply edits to the Ontology without disrupting the main Ontology data. Users can create and compare multiple scenarios and merge edits to main Ontology data.

    Limitations

    Consider the following limitations when building out your workflow using scenarios:

    • Because merging a scenario uses an action, apply action limits apply to merging a scenario.
    • Scenarios are auto-rebased every 10 minutes on top of their base.
    • Scenarios cannot be rebased at will or be configured with rebase intervals.