I had a new deployment in DataRobot using the mode of [external deployment](https://docs.datarobot.com/en/docs/mlops/deployment/deploy-workflows/ext-cus-model-ext-env.html). The model predictions are made externally but monitored by DataRobot MLOps agent.
My deployment was set up with "target monitoring" and "feature drift tracking" enabled. There is also a baseline training dataset.
However, in my dashboard there is no Data Drift information.
I had the prediction stat and data reported using `
Thanks Bogdan for outlining the diagnosis steps. I have actually manged to resolve this issue. It's because one of the columns I passed in as prediction data does not comply with Datarobot requirement. The data drift is working now after I corrected it.
One suggestion for future deveopment is whethere there can be more meaningful error message returned from API calls. My API calls are sent to DataRobot without error, so I was asuuming everything I passed in was correct.
Speaking of the agent log, it's in a special fs_spool format which is not very friendly to be read.
If you can't see service stats, this could be the result of a number of issues: