Changelog
- 0561679 Add stale issue/PR bot workflow (#5061)
- f72fa09 Feat/remove restart policy never if autoscaler v2 (#4816)
- cdfd3ed Fix outdated comment about Kueue's suspend limitations (#5185)
- eb63600 Forward Kubernetes Node infrastructure events to Ray custom resources (#4978)
- 99d1850 History Server: collect rotated Ray logs before overwrite (#5101)
- 0acedf8 K8s 1.37 client go / envtest updates (#5226)
- 39b5d98 K8sJobMode: Honour Ray address overrides in submitter health checks (#5281)
- 2cc93fc Remove manual client go metrics (#5385)
- c073a2d Updating remaining K8s deps and controller-runtime for v1.37 (#5255)
- aac52c9 Was v1alpha2 batch scheduler (#4962)
- 8ce5a28 [APIServer] Wait for the terminate-actor job, not the create-actor job (#5103)
- a647551 [Bug][kubectl-plugin] Check RayJob status after port-forward loss (#5147)
- 8614968 [Bug][kubectl-plugin] Fix log output directory and Ray image defaults (#5053)
- 2e3606f [Bug][kubectl-plugin] Tear down the port-forward child when a session is terminated (#5181)
- 7b41dbc [Buildkite] upload e2e test results to buildkite test suite (#5169)
- 5fd88bb [CI] Grant contents write to workflows (#5428)
- cb2cdc2 [CI] [History Server] Enhance CI ecosystem (#5247)
- dc80c39 [CI] [History Server] temp remove history server e2e (#5342)
- 324126e [CI] add flaky test workflow bot (#5450)
- 4cde56c [CI] make the Test Python Client operator build retry actually retry (#5306)
- 757f55d [CI][History Server] Replace MinIO with RustFS for e2e tests (#5357)
- 06b981f [CI][kubectl-plugin] Deflake session reconnect e2e test (#5106)
- 3ec5cbb [CI][python-client] Publish kuberay-client to PyPI (#5346)
- 215b98f [Doc] Repin ML samples to base ray image with runtime-installed deps (#5194)
- 047da22 [Docs] Fix DeletionStrategy godoc rendering in the generated API reference (#5182)
- 60d070e [Docs] Fix casing and a nonexistent image tag in API doc comments (#5432)
- 8f7820f [Docs] Update contributing docs & AI contribution guidance (#5399)
- 9dc5527 [Docs][Dashboard] Fix the stale dashboard README (#5187)
- 2aab481 [Docs][History Server] Replace MinIO with RustFS in docs (#5384)
- 1d5e26f [Feat][RayCronJob] Include metadata (e.g. labels,annotations) in RayCronJob jobTemplate (#5383)
- fbc5c88 [Feature] Add resource mapping for Huawei Ascend NPUs (#5071)
- 91f1833 [Feature] Support in-place gang resize for Kubernetes WAS v1alpha3 (#5266)
- 38449c9 [Feature][History Server] Make the collector image configurable on the operator (#5207)
- 775e119 [Feature][Volcano] Populate SubGroupPolicy from Ray worker replicas (#5249)
- 078695f [Fix] Unmask CI failures and fix ready-endpoints check in TestZeroDowntimeUpgradeAfterOperatorUpgrade (#5320)
- 22efc76 [Fix][Prometheus] Add missing provisioned condition descriptor to Describe() (#5311)
- b4ea1a5 [Flaky Test] Apiserver flaky fix (#5177)
- e76908b [Flaky] history server e2e test remove port forward (#5307)
- 1d8fef0 [Follow-up] feat: address review comments on GCS active-passive-head API (#5289)
- 6bf05eb [Grafana] Update dashboards to Ray 2.58 (#5259)
- 9b66ecd [Helm][Feat] Expose enableIngress and ingressOptions in the RayCluster chart (#5195)
- f4ac6d1 [Helm][Feat] Expose spec.authOptions and spec.upgradeStrategy in RayCluster Helm Chart (#5197)
- f2973a6 [History Server] Add aliyunoss History Server deployment and proxy head samples (#5156)
- a4e9d63 [History Server] Anchor the GCS GetContent search at the configured root dir (#5077)
- 647e17f [History Server] Auto reload when changing from running to dead cluster (#5467)
- be548d3 [History Server] Bump Ray to 2.58.0 in examples and e2e test data (#5363)
- 8b7a9d6 [History Server] Cache JSON bytes and share the decoded snapshot via a weak pointer (#5285)
- 5914ac0 [History Server] Cache snapshot struct in the session LRU instead of JSON bytes (#5208)
- 5290ece [History Server] Disable live cluster support by default (#5199)
- 0523eda [History Server] Fix dashboard token login with --proxy-server-url (#5472)
- 7ca9b22 [History Server] Refactor get dir helper function (#5216)
- e3653d7 [History Server] Rename GetContent first param clusterId -> prefix (#5068)
- 175e847 [History Server] Replace RayCluster with RayJob sample yaml (#5150)
- 8af0e52 [History Server] Set timeout to 2 minutes for S3 and aliyunoss (#5390)
- 11c0fa2 [History Server] Upload running history server logs by chunk (#5347)
- c1a3a58 [History Server] add reload button in select cluster page (#5463)
- 06ce973 [History Server] create helper to extract similar pattern for session node listing (#5113)
- 4337247 [History Server][CI] add unit tests for the lineage task summary (#5128)
- a6f9d3e [History Server][test] Add unit tests for state_transitions (#5250)
- 3972e1a [RayCluster] [Fix] RayCluster suspend leaks the Volcano PodGroup, leaving queue resources reserved (#5200)
- 9ff7086 [RayCluster][CI] add e2e test for autoscaler scale-up with mTLS (#5066)
- 507a2f5 [RayCluster][GCS FT] Do not inherit head Pod anti-affinity in the Redis cleanup Job (#5448)
- f1db803 [RayJob] Align existing RayCluster sample name (#5170)
- a3c2c1a [RayService] Assert rollback reason in TestReconcileRollbackState (#5076)
- edda320 [RayService] Support attaching incremental upgrade to an existing Gateway (#5018)
- feeaf72 [Sample] Fix samples (#5154)
- 3849337 [Stale bot] disable stale bot for issue (#5335)
- fab7c40 [TAS][Webhook 1/n] Write topology API backbone (#5260)
- 475c0a0 [TAS][Webhook 2/2] Deliver node labels to Ray workers through admission webhooks (#5300)
- 5cc0a27 [TAS][Webhook] Reshape topology.labelMappings into workerGroupSpecs[].labelRefs (#5329)
- af33b14 [Test][History Server] Fix Build historyserver flaky unit test: TestShutdown_DrainsRetriedTasks (#5245)
- a439529 [docs] add CNCF sandbox and LF Projects footer, move GOVERNANCE.md to root (#5370)
- ee4478f [docs] refine the community governance docs (#5254)
- d4e4ca1 [ray-operator] Add spec.idleSuspend and spec.idleTerminationOptions for v1 RayCluster CRD (#5210)
- 1a27a45 [ray-operator] Bump golang.org/x/net, golang.org/x/text and cert-manager to fix CVEs (#5486)
- 9dd658f [ray-operator] add sandbox samples (#5153)
- f94f93b build: add docker build in history server (#5248)
- ff5d9da clarify the security policy and add the report form (#5225)
- ffec815 comply the cncf requirements on the maintainers table (#5212)
- eed5b24 comply the cncf requirements on the security policy (#5219)
- fa76e8c example: Add agent-sandbox suspend/resume example with pod snapshots (#5213)
- fc84673 feat: add API and webhook validation for GCS active-passive-head (#4910)
- 8586d3a feat: add ScaleGate field to ScaleStrategy for fast fallback on quota exceeded (#5094)
- 545d0f1 fix(rayjob): prevent transient cluster name for initially suspended jobs (#5293) (#5296)
- 80068bc fix: remove no-op WorkersToDelete resets in RayCluster controller (#5217)
- 2fdafb2 helm: preserve maxReplicas: 0 in the RayCluster chart (#5481)
- 8e4dee8 helm: support common envFrom for Ray containers (#5034)
- 7c84e51 historyserver: fix in-cluster service discovery URL format (#5351)
- bbc8ac2 main: use resolved config for log encoders (#5322)
- 36a7cdd stop propagating proc.returncode in locust_runner.py. see #5109 (#5117)
- 11d0d32 update KubeRay version to 1.8.0-rc.0 (#5502)
- dc1f48a update the Company/Organization column in PEOPLE.md (#5258)