We have an article on how to deploy a DataRobot scoring code jar model on Databricks, and monitor it from DataRobot MLOps. This leverages an agent architecture that reports performance and feature/prediction data back to DataRobot. You can find it here.
You can additionally monitor non-DataRobot models as well. Note the actual code calls simply send performance stats (number of rows, time to score) and data (features, prediction/model outputs) back to DataRobot, which has an entry created for the deployment based on model characteristics and a training dataset you provide. The model itself is independent from the monitoring operations; any custom model, or model created by another platform, could be monitored.
In general, you can use DataRobot MLOps with models deployed anywhere! DataRobot provides MLOps tracking agents that can be used monitor models where they live, and this info communicated back to the DataRobot MLOps platform, providing data drift, service health, and accuracy monitoring for your model.
A special case is a model deployed to Spark (Hadoop, Databricks, AWS EMR). Provided you could put your predictions and features in a Spark DataFrame, you will be able to monitor the model with ease. Hope this helps!
If you interested, I would encourage you checkout the webinar and articles linked below