Releases: kubeflow/kfp-tekton
KFP-Tekton v1.0.0 with KFP v1.7.0, Tekton v0.27.0
- Tekton v0.27.0 support (#700)
- Rebase KFP core to 1.7.0, KFP SDK to 1.7.2 (#713, #699, #705, #718)
- PipelineLoop and custom tasks now support timeout (#706, #709)
- Support multi-step parameter passing if the Tekton results are too large (#702, #711)
- Any-sequencer, custom task, and unit tests are now based on component.yaml and compatible with all supposed V2 IR compiler features in KFP 1.7.0 release. (#676, #677, #675, #680, #685, #691, #692, #693, #708)
- Enhance Tekton variable sanitization based on respective Tekton Spec. (#686, #688, #689)
KFP-Tekton v0.9.0 with Tekton v0.25.0, Loops Parallelism and Graceful Termination API support
- Tekton 0.25.0 support (#626)
- Parallelism support for Pipeline Loop (#589, #592)
- Inline taskSpec support for Parallel Loop and user custom task definition. (#633, #636, #658)
- Extend KFP-Tekton Run API to create recursive pipeline loop resources (#512, #674)
- Add Tekton 0.25.0 graceful termination API on KFP-Tekton. (#631)
- Enhance any sequencer, recursive loop, and custom task on conditional pipelines. (#587, #635, #650)
- Pipeline loop CLI validation on Pipeline loop resources and inline taskSpec. (#614)
KFP-Tekton v0.8.0 with Graph Recursion and Tekton Custom Task Controller support
- Rebase to KFP 1.5.0 (#555, #513)
- Support graph recursion for KFP on Tekton (#515, #556)
- Support Tekton custom task definition on KFP (#478, #526, #528, #544)
- KFP Caching are enabled by default now and configurable from the server side. (#507, #537, 576)
- Add support for Tekton pipelineRun scope metadata configuration (#487)
- Support Versioning for Tekton loop resources and enhance nested loop usage (#484, #485, #491, #495, #523)
- Support new Tekton pipelineRun status: PipelineRunStopping (#545)
- Add condition and custom task support for any-sequencer (#539, #557, #583)
- Add condition support for custom tasks (#478, #564, #579)
KFP-Tekton v0.7.0 with conditional loops and dynamic parameters support
- Custom task loops that support Dynamic parameters and loops with conditions (#440, #452, #455, #456)
- Tekton 0.21.0 API support for dynamic parameter passing to custom task and finally conditions (#476)
- Preview for Caching support on Tekton (#424)
- Rebased to KFP 1.4.0 (#481)
- Added End to End testing with IBM Cloud (#459, #466, #472)
KFP-Tekton v0.6.0 with Tekton v0.20.1, Any Sequencer support
- New feature: Any sequencer. When any one of the task dependencies complete successfully, the dependent task will be started. (#399, #401, #414)
- Enhance Tekton variables: Properly convert Tekton Context variables as pipeline/task parameters. (#421)
- Rebase to KFP 1.3.0 (#423)
- Bugs fixes on nested condition and added golang tests. (#400, #404, #395, #403)
- Tekton 0.20.1 support. (#428)
KFP-Tekton v0.5.0 with Tekton v0.16.3, whenExpression support
Support Tekton 0.16.3 features (#321)
Migrate Tekton conditions to Tekton whenExpression (#321) (#328)
Support more Tekton context variables such as pipeline UID (#321)
Support task wise metadata on Tekton (#321)
Allow task wise display name annotations (#383)
UI supports both Tekton conditions and whenExpressions (#328) (#373)
Add Safari browser support (#369)
KFP-Tekton with standalone installation, logs API, UI support for retrieving archived logs etc.
Enables support for standalone installation, logging API, UI support for retrieving archived logs, dynamic PVC creation for big data passing, customizable S3 endpoint, better SDK documentation etc. Rebased on KFP v1.0.4
- Add KFP Tekton release deployment yaml (#347)
- Migrated to Kubeflow pipeline v1.0.4 (#337)
- Logging API from the api-server (#299)
- UI able to retrieve the archive logs and the api-server logging API (#305)
- Big data passing now can dynamically create PVC (#309)
- Admin can provide customizable container spec for processing artifacts. (#336)
- Add s3 endpoint instructions for admin use case (#323)
- Updated SDK instructions (#358)
Multi User Kubeflow Pipelines with Tekton, Log Archival and OpenShift support
Enables support for running Kubeflow Pipelines with Tekton in multi-user mode. Additionally enables support for OpenShift with OpenShift Pipelines. Also adds artifact visualization, log archival etc.
Using Kubeflow Pipelines 1.0 and Tekton 0.14.0 for this release.
- Rebase to KFP 1.0 (#251)
- Multi-user support (#259) (#275)
- Available on OpenShift, and enabled with OpenShift Pipelines Tekton engine (#285) (#302)
- Artifact visualization with KFP-Tekton (#266)
- Log archival (#248) (#281)
- Admin configurable artifact, logging, and strip EOF settings (#281)
- Auto Strip EOF newlines for Tekton results (parameter outputs) (#281)
- Can support Tekton 0.15.0 API (#260) (#262)
Kubeflow Pipeline End to End with Tekton backend
Enables support for running Kubeflow Pipelines end to end with Tekton backend. Using Kubeflow Pipelines 0.5.1 and Tekton 0.14.0 for this release.
- Create your Pipeline using Kubeflow Pipelines DSL, and compile it to Tekton YAML.
- Upload the compiled Tekton YAML to KFP engine (API and UI), and run end to end with logging and artifacts tracking enabled.
First release for KFP-Tekton Compiler
The current release provides a Kubeflow Pipeline compiler for Tekton. It can take KFP DSL and compile it to Tekton YAML, which can be run on Tekton directly. Tested with KFP v0.5.1 and Tekton v0.13