Releases: awslabs/kubeflow-manifests
Releases · awslabs/kubeflow-manifests
v1.5.1-aws-b1.0.0
What’s New
This release offers the following features:
- Upgrade Kubeflow components for 1.5.1. Component versions as listed in components versions table
- Access AWS services from Katib using the AWS IAM Roles for Service Accounts (IRSA) integration with Kubeflow Profiles
- Access AWS services from pipeline pods using the AWS IAM Roles for Service Accounts (IRSA) integration with Kubeflow Profiles
- Switch from KFServing to KServe as default serving component. Component guide for model serving over load balancer endpoint using KServe
- AWS optimized Jupyter notebook server images for TensorFlow-2.6.3 and PyTorch-1.11
- Bug fix to remove unused mysql-pod and pv-claim from pipeline component (#222)
Updated documentation available at: https://awslabs.github.io/kubeflow-manifests/release-v1.5.1-aws-b1.0.0/docs/
Known Issues
- #257
- #117 (Workaround documented in issue)
- #118 (Workaround documented in issue)
- kubeflow/pipelines#7361 (Terminating the pipeline run does not trigger the deletion logic programmed via the signal handled in a component. This affects all components in general. Terminate functionality in SageMaker components for Kubeflow pipelines is also affected. Workaround is to manually stop the training jobs)
Full Changelog: https://github.com/awslabs/kubeflow-manifests/commits/v1.5.1-aws-b1.0.0
v1.4.1-aws-b1.0.0
What’s New
New Website : https://awslabs.github.io/kubeflow-manifests/
This release offers the following integrations and deployment options:
- Automated Setup Script for Amazon Relational Database Service (RDS) and Amazon S3
- Automated Setup Script for AWS Cognito
- Automated Setup Script for Amazon Elastic File System (EFS)
- Automated Setup Script for Amazon FSx for Lustre
- Automated Setup Script for exposing Kubeflow over Application Load Balancer
- AWS IAM Roles for Service Accounts (IRSA) integration with Kubeflow Profiles with support for Notebook component
- Component Guide for Kubeflow KServe/KFServing on AWS
- Amazon CloudWatch Container Insights Integration to capture EKS logs and metrics
- Source for Kubeflow Notebook containers is now part of the repo
- [Bug Fix] AWS Cognito Logout
Known Issues
Contributors
- @akartsky, @AlexandreBrown, @goswamig, @judyheflin, @mbaijal, @rrrkharse, @ryansteakley, @surajkota, @wenjinsitu
Full Changelog: https://github.com/awslabs/kubeflow-manifests/commits/v1.4.1-aws-b1.0.0
v1.3.1-aws-b1.0.0
What's New
This release offers the following integrations and deployment options:
- AWS optimized Jupyter notebook server images based on AWS Deep Learning Containers
- Integration with AWS Application Load Balancer to manage external traffic using the AWS Load Balancer Controller
- Integration with AWS Certificate Manager and AWS Cognito for TLS and authentication
- Integration with Amazon Relational Database Service (RDS) in Pipelines and AutoML(Katib) for persistent metadata store
- Integration with Amazon S3 in Pipelines for persistent artifacts store
- Integration Amazon EFS CSI driver to manage Amazon Elastic File System (EFS) as persistent workspace or data volumes
- Integration with Amazon FSx CSI driver to manage Lustre file systems as persistent workspace or data volumes
- Detailed end to end deployment guides for a number of deployment options
- Move to Kustomize based installation (no longer use kfctl which was used in Kubeflow-1.2)
- Compatible with EKS v1.19, v1.20 and v1.21
Known Issues
Contributors
- @akartsky, @AlexandreBrown, @goswamig, @jlbutler, @judyheflin, @mbaijal, @rrrkharse, @ryansteakley, @surajkota
Full Changelog: https://github.com/awslabs/kubeflow-manifests/commits/v1.3.1-aws-b1.0.0
Contributors
goswamig, jlbutler, and 7 other contributors