Aller au contenu
login
arrow_backRetour aux issues
kubeflow/trainer #3120

Automate managing of network policy resources for every train job

ecoDébutant help wanted good first issue kind/feature area/controller

descriptionDescription

### What you would like to be added? As discussed in the wg call this week, creating this issue to track automating network policies for trainjobs to restrict pod-pod communication, egress traffic and access to data cache. It should allow platform owners to optin/optout via trainer config offering out-of-the box experience for ml engineers ### Why is this needed? Providing zero-trust security for jobs in multi-tenant environments ### Love this feature? Give it a 👍 We prioritize the features with most 👍
codeOuvre sur GitHub