Deploy workflows¶

After you've connected an agentic workflow custom model to an agentic playground, compared and tested the workflow, and made any required modifications, you can register the production-ready agentic workflow for deployment to Console.

Register an agentic workflow¶

After experimenting in the agentic playground to build a production-ready agentic workflow, register the custom agentic workflow in the Registry workshop, in preparation for deployment to Console.

To register an agentic workflow from an agentic playground:

1. From an agentic playground, find the register option in the comparison chat, or a single-agent chat:

   Comparison chatSingle-agent chat

   Click the actions menu for the agentic workflow you want to register, then click Register agentic workflow.

   

   Actions menu locations

   The actions menu is always available in the Workflows panel for connected agentic workflows. In addition, when a workflow is selected for comparison, the actions menu is available next to the agentic workflow name in the comparison chat area.

   Click Register agentic workflow in the upper-right corner of the single-agent chat window.

   

   The agentic workflow opens in the Registry Workshop.

2. In Workshop, ensure the agentic workflow is assembled correctly by reviewing the following sections:

   

   • Environment, Dependencies, and Files
   • Evaluation and moderation
   • Runtime parameters
   • Settings

   LLM gateway access runtime parameter requirement

   To use the LLM gateway for an agentic workflow, the ENABLE_LLM_GATEWAY_INFERENCE runtime parameter must be provided in the model-metadata.yaml file and set to true.

3. (Optional) Click Test workflow to provide a test dataset and test the agentic workflow response through the chat and/or score hooks.

4. Click Register a workflow to open the Register a workflow page. Under Configure the workflow, the Target is set based on the workflow you're registering and the Target type is set to Agentic Workflow. Select one of the following registration options:

   • Add a version to an existing registered workflow: Increment the version number and add a new version to the selected registered workflow.

   • Create a new registered workflow: Create a registered workflow and the first version (V1).

     Then, configure the following fields:

     

     Field	Description
     Registered model name / Registered model	Do one of the following: Registered model name: When registering a new workflow, enter a unique and descriptive name for the new registered workflow. If you choose a name that exists anywhere within your organization, a warning appears. Registered model: When saving as a version of an existing workflow, select the existing registered workflow you want to add a new version to.
     Registered version name	Automatically populated with the workflow name, date, and time. Change or modify the name as necessary.
     Registered version	Assigned automatically. This displays the expected version number of the version (e.g., V1, V2, V3) you create. This is always V1 when you select Register as a new model.
     Optional settings
     Registered version description	Enter a description of the business problem this workflow solves, or, more generally, describe the workflow represented by this version.
     Tags	Click + Add tag and enter a Key and a Value for each key-value pair you want to tag the workflow version with. Tags added when registering a new workflow are applied to V1.

     Note

     If you click Cancel on this page to return to the Registry, you lose the configuration progress on this page.

5. Click Register a workflow.

   The agentic workflow version opens on the Registry > Models page with a Building status.

Deploy an agentic workflow¶

Once the agentic workflow is registered (listed in the Models tile in Registry), you can deploy it to Console as you would any other model type, providing access to DataRobot monitoring capabilities.

To deploy an agentic workflow from Registry:

1. On the Registry > Models page, if the registered agentic workflow isn't already open, click the Agentic Workflow tab and locate the workflow to deploy.

   

2. Click the agentic workflow to open it. If it has the Ready for deployment status badge, click Deploy.

   

3. Configure the deployment settings for the agentic workflow. In particular, review the following sections:

   • Prediction history and service health: Ensure the intended DataRobot Serverless prediction environment is selected. For this deployment type, you can only select serverless environments.
   • Data exploration: Ensure that prediction row storage is enabled for data exploration and analysis.
   • Custom metrics: Ensure automatic association ID generation for prediction rows is enabled—and an association ID feature name is provided—to enable custom metrics for the deployment.
   • Runtime parameters: Ensure the runtime parameters provided during agentic workflow assembly are correct, and edit them if necessary.

   Setting configuration post-deployment

   You can configure these settings after the workflow is deployed; however, some settings, like runtime parameters, require temporary deactivation of the deployment.

4. After configuring the deployment settings, click Deploy model.

   

   The Creating deployment dialog box appears. Wait for deployment creation, or click Return to deployments to open Console.

In Console, on the Deployments page, click the Agentic Workflow tab to view a list of agentic workflow deployments.

A fully configured agentic workflow deployment has access to the following monitoring and predictions features: