I believe each project is maintaining some github actions that are either in someone's personal repo or within the project repo that other project teams can leverage. This will also enable the kubeflow community to align on some of the common best practices (clean up github workflows) and benefit from the kubeflow community's workflow ideas.
For e.g. in the kubeflow-pipelines project, I believe we have few github actions that the community can benefit from:
https://github.com/kubeflow/pipelines/tree/master/.github/actions/check-artifact-exists - which validates whether an artifact exists in the list of github artifacts for that repo or not (this can be used to speed up the image build process - for e.g. pipelines use artifacts names by last commit id, so we do not need to rebuild images if we've previously built it with the same commit)
https://github.com/kubeflow/pipelines/tree/master/.github/actions/junit-summary - Read the junit style xmls (you can provide multiple xmls to aggregate test results) and generate a step summary with list of failed test cases if any and you can custom metadata like html report link as well
https://github.com/kubeflow/pipelines/tree/master/.github/actions/create-cluster - Create kind cluster with retry (with any kubernetes version, old to latest)
I believe each project is maintaining some github actions that are either in someone's personal repo or within the project repo that other project teams can leverage. This will also enable the kubeflow community to align on some of the common best practices (clean up github workflows) and benefit from the kubeflow community's workflow ideas.
For e.g. in the kubeflow-pipelines project, I believe we have few github actions that the community can benefit from:
https://github.com/kubeflow/pipelines/tree/master/.github/actions/check-artifact-exists - which validates whether an artifact exists in the list of github artifacts for that repo or not (this can be used to speed up the image build process - for e.g. pipelines use artifacts names by last commit id, so we do not need to rebuild images if we've previously built it with the same commit)
https://github.com/kubeflow/pipelines/tree/master/.github/actions/junit-summary - Read the junit style xmls (you can provide multiple xmls to aggregate test results) and generate a step summary with list of failed test cases if any and you can custom metadata like html report link as well
https://github.com/kubeflow/pipelines/tree/master/.github/actions/create-cluster - Create kind cluster with retry (with any kubernetes version, old to latest)