microsoft / microsoft/dstoolkit-mlops-v2

Add capability to view metrics in graph/chart form at the parent job level.

Open
#152 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
33
Forks
17
PR merge metrics
No merged PRs in 30d

Description


Name About
🚀 Increase Parent job observability As a User, Metrics should be visible from the parent job, so Charts/Graphs can be Created/Viewed.

Problem Statement

Currently metrics are only provided at the child job level. Because of this, charts/graphs can't be created properly to give better insights.

Proposed Solution

Metrics need to be visible at the parent job level to be able to create charts and graphs for the experiment.
Whenever a metrics log call is made, another call could be made with the run_id field populated with the parent's run id.

Example:

mlflow.log_metric("scoring_mse", mse) 
mlflow.log_metrics("scoring_mse", mse,  run_id=parent_run_id)

The following is an example of how to retrieve the parent run id:

current_run_id = active_run().info.run_id
parent_run = get_parent_run(run_id=current_run_id)
parent_run_id = parent_run.info.run_id

Once metrics are populated to the parent job, you can have dashboards like this:
aml-parent-job-metrics

Acceptance Criteria

  • Metrics are propagated to the parent run.
  • Charts/Graphs are populated.
  • A markdown doc is created giving details on the why and how it was done.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No repository files or tests are named; start by tracing the Python MLflow metric-logging path and the active_run/get_parent_run entry points shown in the issue. Determine where parent-run metrics and chart data are handled. Done means metrics propagate to the parent run, charts or graphs are populated, and a markdown document explains the rationale and usage.

Written by the indexing model from the issue text.

Assessment

Tech stack
azure, python
Domain
data-visualization, machine-learning, observability
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.