Willingness to contribute

Yes. I would be willing to contribute this feature with guidance from the MLflow community.

Proposal Summary

It would be useful to be able to rollback to a previous model version in given stage, especially Production.

For example, say that I am working on a particular a model and after several iterations, I decide to move version 3 into Production. Now, after a few more rounds development and tuning, I have what I believe to be a better model in version 7 and I decide that I can move it into Production.

After a while, when scoring on the new model version, my analysts inform me that there is some kind of bug with the new version. In such a situation, I would like to be able to quickly move back to the previous version, which I know was working fine, so that no further predictions will be affected.

Motivation

What is the use case for this feature?

The use case is to be able to quickly rollback to a previous model version in a stage, particularly in Production, that was working fine, without having to remember or search through model versions.

Why is this use case valuable to support for MLflow users in general?

The need to move back to a previous model version is a scenario that I believe anyone can come across.

Why is this use case valuable to support for your project(s) or organization?

I have come across the need to rollback, as a result of bugs that exist in the newer model versions that are transitioned.

Why is it currently difficult to achieve this use case?

At the moment, I am not sure if it is possible to achieve this, outside of searching through all of the models and deciding which one to bring back.

Details

No response

What component(s) does this bug affect?

• area/artifacts: Artifact stores and artifact logging
• area/build: Build and test infrastructure for MLflow
• area/docs: MLflow documentation pages
• area/examples: Example code
• area/model-registry: Model Registry service, APIs, and the fluent client calls for Model Registry
• area/models: MLmodel format, model serialization/deserialization, flavors
• area/pipelines: Pipelines, Pipeline APIs, Pipeline configs, Pipeline Templates
• area/projects: MLproject format, project running backends
• area/scoring: MLflow Model server, model deployment tools, Spark UDFs
• area/server-infra: MLflow Tracking server backend
• area/tracking: Tracking Service, tracking client APIs, autologging

What interface(s) does this bug affect?

• area/uiux: Front-end, user experience, plotting, JavaScript, JavaScript dev server
• area/docker: Docker use across MLflow's components, such as MLflow Projects and MLflow Models
• area/sqlalchemy: Use of SQLAlchemy in the Tracking Service or Model Registry
• area/windows: Windows support

What language(s) does this bug affect?

• language/r: R APIs and clients
• language/java: Java APIs and clients
• language/new: Proposals for new client languages

What integration(s) does this bug affect?

• integrations/azure: Azure and Azure ML integrations
• integrations/sagemaker: SageMaker integrations
• integrations/databricks: Databricks integrations