Get mlflow to work with with custom Databricks docker container compute

[tmueller]

We need conda for our python things, so i set up a Compute unit with a custom dockerfile. Now one of our ki engeniers tried to get mlflow working but it seems not connected to the mlflow of Databricks. How can i get mlflow in my container to work with mlflow in databricks?

[Anonymous]

Hi @Tobias Müller​ 

Great to meet you, and thanks for your question! 

Let's see if your peers in the community have an answer to your question first. Or else bricksters will get back to you soon.

Thanks.

[Thijs]

We have this same issues. Did you ever manage to solve this?

[tmueller]

I found the solution after talking to support.

The solution is not in the databricks documentation but in the mlflow

https://mlflow.org/docs/latest/quickstart.html?highlight=databricks#use-mlflow-with-a-databricks-wor...

after you set tracking uri to databricks it worked for us 😄

https://www.mlflow.org/docs/latest/python_api/mlflow.html#mlflow.set_tracking_uri this was the link they send me originaly

[Thijs]

Thanks @tmueller, that indeed solves it!