feat: add helm chart for training-operator

[vicoooo26]

What this PR does / why we need it:

Add a Helm chart for deploying the training-operator on Kubernetes. The chart includes all necessary CRDs (TFJob, PyTorchJob, JAXJob, MPIJob, PaddleJob, XGBoostJob), RBAC resources, webhook configuration, and configurable values for production deployments.

Which issue(s) this PR fixes (optional, in Fixes #<issue number>, #<issue number>, ... format, will close the issue(s) when PR gets merged):

Checklist:

• Docs included if any changes are user facing

[github-actions]

🎉 Welcome to the Kubeflow Trainer! 🎉

Thanks for opening your first PR! We're happy to have you as part of our community 🚀

Here's what happens next:

• If you haven't already, please check out our Contributing Guide for repo-specific guidelines and the Kubeflow Contributor Guide for general community standards.
• Our team will review your PR soon! cc @kubeflow/kubeflow-trainer-team

Join the community:

• Slack: Join our #kubeflow-trainer Slack channel.
• Meetings: Attend the Kubeflow AutoML and Training Working Group bi-weekly meetings.

Feel free to ask questions in the comments if you need any help or clarification!
Thanks again for contributing to Kubeflow! 🙏

[cheyang]

Early feedback (still a draft). The templates are in good shape — the fail guard on replicas>1 without leader election, the probes, the security context, and values.schema.json (incl. the min/maxUnavailable mutual-exclusion) are all nice; helm lint is clean and helm template renders. A few things stood out, mostly from a first-run user’s perspective:

• The PR description says “cert-manager integration”, but the chart doesn’t use cert-manager (no Certificate/Issuer, no caBundle / inject annotation). It relies on the operator’s built-in self-signed cert written into an empty Secret and patching the webhook caBundle at runtime (hence the validatingwebhookconfigurations: update grant). Worth correcting the description.
• The README install commands reference the aliyun-incubator/training-operator repo, but there’s no helm repo add aliyun-incubator <url> step — the documented install won’t work as written.
• Minor: CRDs under crds/ aren’t upgraded by Helm; a note in the README about applying CRD updates manually would help.

[Copilot]

Pull request overview

This PR introduces a Helm chart to deploy the Kubeflow Training Operator on Kubernetes (including RBAC, webhook wiring, deployment/service, and chart docs), and updates CI/test configs to keep Python tests stable with newer kubernetes client changes.

Changes:

• Add a new charts/training-operator Helm chart with templates, values, schema, and README.
• Adjust Python/CI test configurations (e.g., pin kubernetes<36, tweak JAXJob e2e resource limits).
• Add Helm chart packaging/support files (e.g., .helmignore, logo, chart metadata).

Reviewed changes

Copilot reviewed 25 out of 32 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
sdk/python/test/e2e/test_e2e_jaxjob.py	Increases JAXJob e2e container memory limit.
charts/training-operator/Chart.yaml	Adds Helm chart metadata for training-operator.
charts/training-operator/values.yaml	Provides default, configurable chart values.
charts/training-operator/values.schema.json	Adds JSON schema validation for values.
charts/training-operator/README.md	Documents installation, configuration, and values.
charts/training-operator/logo.svg	Adds chart icon asset.
charts/training-operator/.helmignore	Adds Helm ignore patterns for packaging.
charts/training-operator/templates/_helpers.tpl	Adds naming/label helper templates.
charts/training-operator/templates/deployment.yaml	Defines operator Deployment, probes, args, volumes, HA behavior.
charts/training-operator/templates/service.yaml	Exposes metrics and webhook ports via Service.
charts/training-operator/templates/serviceaccount.yaml	Creates ServiceAccount (optionally) for the operator.
charts/training-operator/templates/clusterrole.yaml	Defines cluster-scoped RBAC for the operator.
charts/training-operator/templates/clusterrolebinding.yaml	Binds ServiceAccount to ClusterRole.
charts/training-operator/templates/role.yaml	Defines namespace-scoped RBAC (leases/secrets).
charts/training-operator/templates/rolebinding.yaml	Binds ServiceAccount to Role in the release namespace.
charts/training-operator/templates/validatingwebhookconfiguration.yaml	Creates ValidatingWebhookConfiguration for Training Operator CRDs.
charts/training-operator/templates/secret-webhook-cert.yaml	Creates webhook cert Secret placeholder for cert rotator.
charts/training-operator/templates/configmap.yaml	Optionally provides custom PyTorch init container template ConfigMap.
charts/training-operator/templates/pdb.yaml	Adds optional PodDisruptionBudget for HA deployments.
charts/training-operator/templates/NOTES.txt	Adds post-install guidance notes.
charts/training-operator/templates/tests/test-connection.yaml	Adds Helm test pod to check metrics endpoint connectivity.
.pre-commit-config.yaml	Excludes Helm templates from generic YAML checks.
.github/workflows/test-python.yaml	Pins Python kubernetes<36 for unit-test compatibility.
.github/workflows/test-example-notebooks.yaml	Pins kubernetes<36 and fastjsonschema<2.22 for notebook runs.
.github/workflows/integration-tests.yaml	Pins kubernetes<36 for integration tests.
.github/workflows/e2e-test-train-api.yaml	Pins kubernetes<36 for e2e train API tests.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.