sky-uk/kfp-operator

Issues

• Vertex AI Event Source

  #169 opened 2 years ago
  0

• Manage Vertex AI run configurations

  #168 opened 2 years ago
  0

• Add error checking for missing pipeline versions when submitting KFP jobs (run configurations)

  #166 opened 2 years ago
  0

• Preserve duplicate resource fields between supported version conversions

  #161 opened 2 years ago
  0

• Python 3.9 Support

  #160 opened 2 years ago
  1

• Flaky resource conversion test

  #159 opened 5 months ago
  2

• Controller does not start due webhook race condition

  #156 opened 2 years ago
  0

• Allow to specify Service Account per resource

  #153 opened 5 months ago
  1

• Manage Vertex AI pipelines

  #150 opened 2 years ago
  2

• Downgrade resource versions to v1alpha1

  #146 opened 2 years ago
  0

• React to KFP Kubernetes resources instead of Argo workflows

  #145 opened 2 years ago
  0

• Website background fails to load

  #141 opened 2 years ago
  0

• High Availability

  #138 opened 2 years ago
  0

• Fetched resources do not contain GroupVersionKind information

  #137 opened 2 years ago
  0

• RunCompletion EventSource does not emit events for one-off runs

  #129 opened 2 years ago
  1

• Deleting a non-existing resource should succeed

  #128 opened 4 months ago
  2

• Update documentation for public release

  #120 opened 2 years ago
  0

• Provide fallback pipeline name for scheduled pipeline runs

  #119 opened 2 years ago
  0

• Handle resource updates if runconfiguration doesn't exist in KFP

  #118 opened 2 years ago
  0

• KFP-SDK failures not propagated

  #117 opened 2 years ago
  0

• Update RunConfiguration on Pipeline changes

  #112 opened 2 years ago
  0

• Expose model training metrics

  #109 opened 21 days ago
  1

• Rename ownership labels in workflows

  #108 opened 2 years ago
  0

• Leave workflow deletion to Argo mechanisms

  #107 opened 2 years ago
  0

• Undeleted succeeded workflows succeed all future operations

  #105 opened 3 years ago
  0

• Implement observedGeneration best practices

  #101 opened 3 years ago
  0

• Support pipeline re-runs in Model Update Eventsource

  #100 opened 3 years ago
  0

• Eventsource server returns no servingModelArtifacts for failed pipelines

  #97 opened 3 years ago
  0

• servingModelArtifacts null when no models pushed

  #95 opened 3 years ago
  0

• Fall back to entrypoint as pipeline name for scheduled pipeline runs

  #93 opened 2 years ago
  0

• Run Completion Event contains serving locations even when model was not pushed

  #91 opened 3 years ago
  0

• Rework Argo Workflows into Workflow Templates

  #90 opened 2 years ago
  0

• Support concurrent EventSource clients

  #89 opened 2 years ago
  0

• Expose node tolerations on all pods

  #88 opened 2 years ago
  1

• Upgrade Ginkgo to v2

  #86 opened 3 years ago
  0

• Run completion events are incorrectly serialised

  #82 opened 3 years ago
  0

• Make Operator metrics available from Helm Chart

  #77 opened 3 years ago
  0

• Emit Argo Events for failed training runs

  #75 opened 3 years ago
  0

• Emit Kubernetes events for custom resources

  #74 opened 3 years ago
  0

• Remove kfp-ext

  #73 opened 2 years ago
  0

• Provide information in pipeline update events that identify the pipeline's version

  #68 opened 5 months ago
  1

• Support fixed pipeline versions in RunConfiguration

  #67 opened 2 years ago
  0

• Use non-root user in custom Dockerfiles

  #65 opened 4 months ago
  0

• Manage one-off pipeline runs declaratively

  #64 opened 2 years ago
  0

• Run Python tests on CI

  #54 opened 3 years ago
  0

• Access control

  #50 opened 5 months ago
  1

• Provide short names for CRDs

  #49 opened 2 years ago
  0

• Publish public Helm chart

  #48 opened 2 years ago
  0

• Deleting a RunConfiguration without an ID hangs in Deleting

  #46 opened 3 years ago
  0

• Make operator manager resources configurable

  #45 opened 3 years ago
  0