Submitting jobs in Qubernetes

How it works

You can now submit jobs to be executed without waiting for their completion. This is particularly useful for long-running tasks or when you want to offload work to the cluster while continuing with other activities.

To submit a job, use the --submit flag with your existing q8sctl execute command:

q8sctl execute \
    # other-flags
    --submit \
    app.py

Tracking submitted jobs

It is highly recommended to use an experiment tracking solution (such as MLflow) to keep track of your submitted jobs, their parameters, and results.

You can configure your Qubernetes project to log to MLflow by adding the tracking server URI in the .env.q8s file in the root of your project:

MLFLOW_TRACKING_URI=http://your-mlflow-tracking-server:5000