Operate first Jupyterhub application provides a Jupyter notebook environment for python development and running machine learning workloads. In this repository, we analyze the data generated from the application hosted on the Operate First cluster on Openshift.
For a smooth transition to other namespaces and clusters, we should add the user editable variables in the beginning of the notebook. Some of them include:
A flag for hashing
Namespace name
Prometheus instance URL
Mean or max value during sampling
Resource: Cpu or memory
Acceptance
[ ] Updated notebook with the variables specified at the top.
For a smooth transition to other namespaces and clusters, we should add the user editable variables in the beginning of the notebook. Some of them include:
Acceptance