Realtime production

Realtime production is used to specify the optimization model configuration using the different activities and deploy the model as a REST API for real-time inferences.

Select the Realtime Production tab to access the real-time production quest canvas. This tab is active when you have exported the one of the most favorable optimize model activity configurations from the design quest that you intend to use in the final endpoint deployment process.

After the realtime production quest is created, you must configure the custom endpoint schema before it can be deployed as an endpoint.

When a production quest is created, the same datasets applied in the design quest are automatically copied over. Additionally, all related activity configurations from the design quest are replicated in the production quest to maintain consistency with the selected optimization setup. Note that:

  • Activities configuration cannot be modified, except for the endpoint schema configuration available within the Optimize Model configuration.
  • Data preprocessing supports two transformation activities: Execute SQL and Scripting.
  • Data post-processing is not supported.

Follow these steps to configure the endpoint schema:

  1. Open the existing optimize model activity of the production quest.
  2. Click Add.
  3. Select the parameter, sets, indices, and constants. At least one parameter must be selected. Only one prepared dataset is supported along with multiple master datasets. Sets, indices and constants are optional.
  4. Generate the JSON schema and then save it. The JSON schema can be copied or downloaded to prepare the runtime data and supply to the endpoint.
  5. Click Save.
  6. Click Run.
Note: For custom algorithms:
  • You must select all the parameters used in the algorithm code. The entire input data must be supplied at runtime.
  • The realtime endpoint supports a single output file.
  • Data preprocessing step in the design quest does not support custom algorithm models at all.