When accessing the mlflow from outside the cluster the S3 storage is not exposed, the default mlflow library in python requires flag --serve-artifacts to use tracking server to upload artefacts instead of directly uploading them to the object storage.
From mlflow documentation:
If the MLflow server is not configured with the --serve-artifacts option, the client directly pushes artifacts to the artifact store. It does not proxy these through the tracking server by default.
For data security its better to not expose the S3 storage to the world if possible.
What needs to get done
add config option
if server-artifacts == True, add flag --serve-artifacts to mlflow server execution
Definition of Done
There is a way to enable the serve-artifacts flag for MLflow.
Context
When accessing the mlflow from outside the cluster the S3 storage is not exposed, the default mlflow library in python requires flag
--serve-artifacts
to use tracking server to upload artefacts instead of directly uploading them to the object storage.From mlflow documentation:
For data security its better to not expose the S3 storage to the world if possible.
What needs to get done
--serve-artifacts
to mlflow server executionDefinition of Done
There is a way to enable the serve-artifacts flag for MLflow.