Skip to content

Releases: opendatahub-io/odh-dashboard

v3.6.0-ea1-odh

Choose a tag to compare

@ppadti ppadti released this 19 Aug 13:53
da9cc5b

Notable changes

  • Ongoing features:
    • Admin Setup of LLMInferenceServiceConfig
    • MaaS Admin Settings
    • OGX (Llama Stack) must apply configuration changes without requiring a restart in Gen AI Studio
    • Integrate Kueue Workload Scheduling Information in Model Deployments Overview
    • Feature Store Admin UI
    • Admin Setup of LLMInferenceServiceConfig
    • Include NIM in new model serving wizard UI
    • MaaS Consumer Portal
    • Model Type and Capability Metadata Schema with API Filtering
    • Surface Tool Calling Configuration During Deployment (Stage 2)
    • MCP catalog and registry integration
  • Miscellaneous bugs and improvements

What's Changed

  • Update Tekton files to version v3.0.1-odh by @odh-devops-app[bot] in #5465
  • Upversion dashboard by @ppadti in #5467
  • Update Tekton files to version v3.0.2-odh by @odh-devops-app[bot] in #5738
  • Upversion dashboard by @manaswinidas in #5741
  • Upversion dashboard by @pnaik1 in #5961
  • Red Hat Konflux update mod-arch-maas by @red-hat-konflux[bot] in #5991
  • update tekton push file for odh by @alexxfan in #5992
  • Upversion dashboard by @manaswinidas in #6204
  • Update mod-arch-maas-push.yaml to not use central pipeline for odh release builds by @MohammadiIram in #6222
  • Upversion dashboard by @ppadti in #6619
  • Upversion Dashboard by @manaswinidas in #7070
  • Upversion Dashboard by @manaswinidas in #7454
  • Release: Tekton pipelines → v3.4.3-odh by @odh-devops-app[bot] in #7528
  • fix the pipeline name by @MohammadiIram in #7541
  • Release: Tekton pipelines → v3.4.3-odh by @odh-devops-app[bot] in #7546
  • Refactor on-cel-expression condition in Tekton YAML by @MohammadiIram in #7551
  • Update CEL expression for push event handling by @MohammadiIram in #7552
  • Refactor CEL expression for release push event by @MohammadiIram in #7553
  • Simplify on-cel expression in MLflow release push by @MohammadiIram in #7554
  • Upversion Dashboard by @ppadti in #8079
  • Update the target of the image tag to the branch name by @andrewballantyne in #8097
  • Upversion Dashboard to v3.5.0 by @Taj010 in #8794
  • Feat(RHOAIENG-78072): Kueue integration with Workbenches Microcopy by @claudialphonse78 in #8765
  • Fix(RHOAIENG-78948): lent case removal by @claudialphonse78 in #8803
  • RHOAIENG-65104: Define a stable public API surface for model-serving and eliminate deep component imports by @paulovmr in #8600
  • RHOAIENG-72366: Move template hooks to @odh-dashboard/model-serving/shared by @paulovmr in #8715
  • docs: update distribution rules to reflect workspace membership by @caponetto in #8797
  • fix(RHOAIENG-78902): derive duplicate k8s name from display name so it auto-updates by @mturley in #8807
  • RHOAIENG-75489: Move LabeledConnection type to @odh-dashboard/model-serving/shared by @paulovmr in #8783
  • feat(RHOAIENG-65097): Lift ProjectSelector components to ui-core (Phase 3) by @jenny-s51 in #8755
  • fix(llmd-serving): fix routing and topology config edge cases in deploy wizard edit flow by @mturley in #8764
  • RHOAIENG-79260: Apply go fmt to agent-ops BFF source files by @Gkrumbach07 in #8854
  • feat(dashboard-operator): preserve user-configured replicas and resources per ADR-0005 [RHOAIENG-78745] by @lucferbux in #8846
  • feat(ci): add sidecar module build validation to Konflux simulator by @antowaddle in #8776
  • fix(RHOAIENG-79268): target TypeaheadSelect toggle in E2E connection type dropdown by @DaoDaoNoCode in #8859
  • feat(RHOAIENG-65097): migrate package imports from internal to core packages (Phase 2) by @jenny-s51 in #8723
  • ignore all dependabot patternfly updates by @christianvogt in #8873
  • ci: add kserve backend to rhaii-tilt by @andyatmiami in #8812
  • chore(deps): bump dompurify from 3.4.11 to 3.4.12 by @dependabot[bot] in #8860
  • chore(deps-dev): bump svgo from 2.8.2 to 2.8.3 by @dependabot[bot] in #8861
  • test: Add E2E test for llm-d routing configurations admin page by @YuliaKrimerman in #8778
  • feat(RHOAIENG-65102): move metrics page components to model-serving package by @jenny-s51 in #8728
  • refactor: add host-api bridge and decouple model-serving from more internal imports by @caponetto in #8789
  • feat(RHOAIENG-78901): Show fast and unsupported labels in deployed models table by @mturley in #8810
  • chore: add paulovmr and caponetto to general-approvers by @andyatmiami in #8930
  • chore(RHOAIENG-78547): Add llmdTemplates feature flag to OdhDashboardConfig CRD by @YuliaKrimerman in #8855
  • test(RHOAIENG-76145): reduce CI mock test wall time via test conversion and file splitting by @antowaddle in #8773
  • fixSelect default deployment method for generative models in wizard by @YuliaKrimerman in #8878
  • Remove 'platform.opendatahub.io/part-of' label when creating admin llmiscfg by @emilys314 in #8890
  • RHOAIENG-79266: Fix llmd wizard page UX signoffs bugs and issues by @jira-autofix[bot] in #8871
  • fix(e2e): prevent MLflow test failures on re-runs from stale name collisions by @nananosirova in #8870
  • Remove mlflowPipelines feature flag by @nananosirova in #8941
  • fix(RHOAIENG-62527): grant non-admin RBAC for Ray dashboard URL resolution by @kanishka-commits in #8741
  • UX Changes for the Guided Tours by @Philip-Carneiro in #8929
  • fix(RHOAIENG-79525): add missing selector label to RHOAI kustomization commonLabels by @lucferbux in #8950
  • refactor(gen-ai): deduplicate BeforeEach setup blocks in BFF test files by @gmenher in #8774
  • Fix(eval-hub): Update Security Insights table label size and value casing by @ikeola13 in #8949
  • RHOAIENG-46527: fix URL query parameter encoding in MakeRequest by @gmenher in #8747
  • Sync from kubeflow/model-registry afcc818 by @ppadti in #8947
  • RHOAIENG-78980: Keep MaaS BFF up when maas-api discovery fails by @Griffin-Sullivan in #8804
  • fix(gen-ai): add egress rule for OGXServer port 8321 in network policy by @jharan1 in #8948
  • chore(gen-ai): RHOAIENG-66815 Add some .env and Makefile updates by @jharan1 in #7887
  • feat(RHOAIENG-78652): add ProjectsContextProvider for RHAII distribution by @jenny-s51 in #8780
  • RHOAIENG-38993: Remove static SA token management, use dynamic OCP JWT only by @gmenher in #8364
  • Migrate frontend bundler from webpack to rspack by @christianvogt in #8868
  • fix(autorag,automl): add egress rules for DSPA and OGX ports in network policies by @jefho-rh in #8971
  • feat: Decouple model-registry dependency from model-serving by @jenny-s51 in #8749
  • fix(automl): block non-ASC...
Read more

v3.5.0

Choose a tag to compare

@Taj010 Taj010 released this 27 Jul 17:38
9a27ea7

What's Changed

  • fix(e2e): shorten model registry test fixture prefixes to avoid DNS limit by @mturley in #7188
  • fix: flaky argo workflow and v1 pipeline import tests in pipelines.cy.ts by @caponetto in #7252
  • Feat(RHOAIENG-56731): type guards strengthening by @claudialphonse78 in #7204
  • Fix(RHOAIENG-58437):Text overflow issue for subscription dropdown in create api modal by @claudialphonse78 in #7273
  • feat(mlflow): add BFF status endpoint and DSPA integration by @caponetto in #7119
  • Fix subscription view page when model refs do not exist by @katieperry4 in #7277
  • fix(model-serving): gate AAA checkbox behind genAiStudio feature flag by @Taj010 in #6926
  • Add new general approvers by @manaswinidas in #7278
  • Update oncel exp by @MohammadiIram in #7227
  • feat: Implement BFF inter-communication by @lucferbux in #6404
  • fix(kueue): migrate Kueue API from v1beta1 to v1beta2 by @PR3MM in #7271
  • feat(automl,autorag): support namespace persistence by @MatthewAThompson in #7286
  • Add lint, type-check, and CI workflows for model-registry frontend and BFF by @ppadti in #7122
  • e2e test for vLLM on MaaS by @ashley-o0o in #7004
  • Hide API Key Behind a Toggle by @katieperry4 in #7287
  • chore(automl,autorag): simplify local development set-up when running in federated mode by @chrjones-rh in #7281
  • chore(deps): bump dompurify from 3.3.0 to 3.4.0 in /packages/autorag/frontend by @dependabot[bot] in #7290
  • fix(maas): subscription filter matches displayName and description by @dpanshug in #7270
  • fix(autox): fix 500 error after many duplicate file uploads by @daniduong in #7220
  • Remove api keys empty state and fix status labels by @Griffin-Sullivan in #7300
  • Sort Subscriptions and Policies by Display Name by @katieperry4 in #7294
  • MaaS Microcopy Updates by @ashley-o0o in #7279
  • update jira triage to incorporate observability and perses by @christianvogt in #7311
  • feat: add review agent skill for UI convention compliance by @jenny-s51 in #7064
  • fix(bff): remove unnecessary RequireAccessToService from K8s-only routes (RHOAIENG-55748) by @liday-rh in #7086
  • Sync from kubeflow/model-registry d941024 by @manaswinidas in #7301
  • feat(automl): add One vs Rest view to confusion matrix by @jefho-rh in #7240
  • fix(cypress): fix access mode verification in verifyStorageClassConfig by @srtanish1992 in #7319
  • feat(autorag/automl): add support for stopping and retrying a run by @nickmazzi in #7288
  • Fix(RHOAIENG-58427):Added toolbar with search based on policy name and description by @claudialphonse78 in #7268
  • refactor(autorag): drop openai-go SDK, use typed LlamaStack structs by @jefho-rh in #7283
  • Fix(RHOAIENG-57631):added bff changes for preserving unmanaged field in Policies and subscription by @claudialphonse78 in #7257
  • feat(autox): support selecting files by clicking rows in file explorer by @daniduong in #7114
  • fix(hardware-profiles): add character limits to name and description fields by @kanishka-commits in #6960
  • fix: Use full page navigation for OIDC logout instead of AJAX fetch by @GowthamShanmugam in #7083
  • update inference service check by @sridarna in #7323
  • [Cypress-e2e] add wait for model-catalog pod ready before performance filter test by @ConorOM1 in #7330
  • Fix button order and standardize help icons in Deploy MCP Server modal by @manaswinidas in #7318
  • test: Update Feast tests to avoid feast apply error by @Srihari1192 in #7324
  • fix(automl,autorag): extend dynamic port-forwarding to S3 and LlamaStack paths by @chrjones-rh in #7310
  • home page task assistant by @christianvogt in #7285
  • fix: remove .github exclusion from CodeRabbit path filters by @christianvogt in #7344
  • chore(autorag+automl): improvements - S3 timeout error handling + FileExplorer UX by @GAUNSD in #7219
  • Add agent rule to use PR template when creating pull requests by @christianvogt in #7341
  • chore(deps): bump github.com/moby/spdystream from 0.5.0 to 0.5.1 in /packages/automl/bff by @dependabot[bot] in #7315
  • Evalhub UI updates by @ikeola13 in #7331
  • Sync from kubeflow/model-registry 6e18d5a by @ppadti in #7348
  • update the ux changes by @pnaik1 in #7346
  • fix(model-registry): surface K8s name validation errors on MCP deploy name field by @manaswinidas in #7192
  • feat(autorag/automl): add retry and stop actions to runs table rows by @nickmazzi in #7321
  • chore(deps): bump dompurify from 3.3.3 to 3.4.1 in /packages/automl/frontend by @dependabot[bot] in #7359
  • chore(automl,autorag): refresh embedded pipeline YAMLs from upstream by @chrjones-rh in #7307
  • Fix(RHOAIENG-58940) :Added fix for pause to be disabled for clusterselector ray job by @claudialphonse78 in #7349
  • [Cypress-e2e] add polling step for validated model cards by @ConorOM1 in #7367
  • Add e2e tests for MLflow experiments page by @nananosirova in #7289
  • Disabled actions for deleting subs and policies by @katieperry4 in #7309
  • fix(model-serving): add optional chaining for ServingRuntime spec access by @srtanish1992 in #6855
  • fix(e2e): handle other catalog sources when verifying disable in testSourceEnableDisable by @manaswinidas in #6734
  • [Cypress-e2e] Add model registry custom properties retention test by @ConorOM1 in #7313
  • fix(model-registry): scope in-cluster K8s client to SAR-only interface in MCP deployment auth by @manaswinidas in #7259
  • fix(storage): correct storage percentage for shared and block storage by @DaoDaoNoCode in #6829
  • Add mock tests for mlflow experiments and prompts pages by @nananosirova in #7371
  • test: prevent false-positive duplicate-name check in argo workflow version test by @caponetto in #7385
  • Fix dev proxy auth header forwarding in federated modules by @mturley in #7065
  • RHOAIENG-56285: [AutoRAG] Remove context_correctness as the metric available for optimization by @jira-autofix[bot] in #7389
  • fix(cypress): stabilize prompt management E2E test polling by @antowaddle in #7393
  • RHOAIENG-56285: [AutoRAG] Remove context_correctness as the metric available for optimization by @jira-autofix[bot] in #7398
  • chore(automl,autorag): configure page css -> scss by @MatthewAThompson in #7356
  • feat(autorag): filter system namespaces from project selector by @MatthewAThompson in #7380
  • fix(network): simplify NetworkPolicy egress to unrestricted in-cluster by @lucferbux in #7399
    *...
Read more

v3.5.0-odh

Choose a tag to compare

@Taj010 Taj010 released this 24 Jul 20:36
0308a02

Notable Changes

  • New features:
    • Agent deployment wizard with end-to-end observability
    • Agent configuration saving (guardrails, MCPs, prompts)
    • MCP Catalog Admin UI with granular access controls
    • MCP support level labels (partner vs community)
    • Agent Catalog for discovering and deploying agent starter kits
    • Multimodal support in Gen AI Studio Playground (image and audio)
    • AI security insights and risk assessments in Model Catalog
    • RBAC role creation UI for custom namespace-scoped roles
    • Inline connection validation with instant error feedback
    • Adding existing secrets support
    • LLMD deployment templates in the deployment wizard
    • GPU topology dashboard for accelerator usage
    • Auto RAG in-UI dataset creation and direct pattern testing
    • Guided tours for new features during upgrades
    • Homepage refactor with personalized task assistant
    • Starter prompts loadable from registry for org-wide sharing
  • Ongoing features:
    • AutoML visual pipeline layout and advanced evaluation visualizations
    • Feature Store & Workbenches bidirectional discoverability
    • MaaS model endpoint registration (internal and external)
    • MaaS admin settings reorganization under subscription management
    • Queue workload visibility (preempted, evicted, admitted status)
    • Perses dashboards moving from Tech Preview to General Availability
    • Model observability and tracing with live metrics and MLFlow trace links
    • SKS foundation enabling modular distributions across Kubernetes
    • Default vector store migration to PostgreSQL
  • Miscellaneous bugs and improvements

What's Changed

  • feat: implement AgentProfile ConfigMap storage and POST endpoint by @NickGagan in #7917
  • Rhoaieng 64748 add description to models in api key modal by @sridarna in #7778
  • fix(e2e): fail fast when deployment grid shows Failed status by @FedeAlonso in #7911
  • chore: add OWNERS entry for dashboard-operator directory by @lucferbux in #7925
  • feat(dashboard-operator): migrate CRD API group and add platform contract by @lucferbux in #7795
  • RHOAIENG-36320: use KServe status.addresses for LLMInferenceService endpoint by @gmenher in #7882
  • feat(model-serving): add conditional "Preconfigure deployment" wizard step by @mturley in #7763
  • support prompt variables in view code by @danpierce1 in #7938
  • feat: add config endpoints and connection-types handler to Core BFF by @caponetto in #7897
  • docs: document Perses dashboard contribution workflow for component teams by @christianvogt in #7918
  • fix(security): update systeminformation to 5.31.6 [RHOAIENG-67429] by @jira-autofix[bot] in #7929
  • Add Start stop for NIMService extension by @emilys314 in #7922
  • Evaluation run configuration validation and primary metric selection by @ikeola13 in #7894
  • RHOAIENG-62254: BFF audio transcription endpoint and generalized media upload by @Lucifergene in #7808
  • Add Konflux build simulation validation by @antowaddle in #7425
  • Fix(RHOAIENG-66732):resolved the feature store token issue by @claudialphonse78 in #7883
  • RHOAIENG-62255, RHOAIENG-62257, RHOAIENG-62256: Frontend ASR model discovery, selector, and audio transcription UX by @Lucifergene in #7898
  • add mock cypress test for evalhub compare by @pnaik1 in #7931
  • RHOAIENG-50649: add total_count and substring filter to prompts list by @gmenher in #7910
  • RHOAIENG-66939, RHOAIENG-66940: View Code snippets for ASR and Vision multimodal features by @Lucifergene in #7913
  • Add key count to BFF and update label in frontend by @Griffin-Sullivan in #7927
  • feat(gen-ai): implement AgentProfile CRUD API (List, Get, Update, Delete) by @NickGagan in #7941
  • feat(automl): Update pipeline YAMLs with eval_metric parameter by @nickmazzi in #7930
  • feat(dashboard-operator): dynamic dependency discovery, ConfigMap flags, and architecture docs by @lucferbux in #7797
  • Add fast-forward sync from odh-release to rhoai-release by @manaswinidas in #7902
  • fix(e2e): clean up stale MaaS resources before retry in testMaaSSubscriptions by @FedeAlonso in #7945
  • Add model serving Ci tags by @emilys314 in #7880
  • fix(observability): explicitly name cluster datasource in Perses dash… by @rsun19 in #7942
  • feat(autorag): add data-testid attributes for E2E test selectors by @jefho-rh in #7957
  • Add sorting to API key table in subscription details by @ashley-o0o in #7950
  • feat(gen-ai): AgentProfile serialize/deserialize and URL-param load by @NickGagan in #7959
  • Chore:Added feature store port forwarding script by @claudialphonse78 in #7947
  • feat(gen-ai): switch View Code snippets from ogx-client to OpenAI SDK by @Lucifergene in #7961
  • Enable Model Training e2e CI tags for automatic test detection by @dpanshug in #7909
  • RHOAIENG-62431 : Add "Connected feature stores" section to Workbench table expanded row by @kanishka-commits in #7932
  • RHOAIENG-63155: Implement permissions rules section with Add Rule modal by @srtanish1992 in #7943
  • feat(gen-ai): gate agent profile Save As button behind agentProfileManagement dev flag by @NickGagan in #7981
  • feat(automl): add eval_metric selection to AutoML configure and results by @jefho-rh in #7921
  • RHOAIENG-68244: Cypress e2e CI fails with File name too long when auto-detected tags are used as results directory name by @paulovmr in #7978
  • Sync from kubeflow/model-registry 3e8c430 by @ppadti in #7979
  • Add connected workbenches modal to Feature Store by @PR3MM in #7924
  • fix(ci): make Konflux simulator actually test the PR image and assert… by @antowaddle in #7962
  • RHOAIENG-68443: Increase Cypress E2E test job timeout from 30 to 60 minutes by @paulovmr in #8027
  • feat(gen-ai): add Agent Profiles tab to AI Asset Endpoints page by @NickGagan in #7987
  • RHOAIENG-68443: Increase Cypress E2E test job timeout from 60 to 90 minutes by @paulovmr in #8072
  • Refactor model transfer jobs handler to use downstream repository layer by @ppadti in #7578
  • add event tracking for side by side comparison by @pnaik1 in #8080
  • dataloader update to accommodate s390X yaml files by @diwakarsharma2 in #8076
  • feat(nim-serving): add form-data extractors and PVC apply/extract for NIM edit wizard by @YuliaKrimerman in #7764
  • fix(cypress): ignore 'Unexpected token ?' uncaught exceptions in E2E … by @antowaddle in #8082
  • Revert "dataloader update to accommodate s390X yaml files" by @sridarna in #8092
  • RHOAIENG-62249, RHOAIENG-62250: Model capabilities annotation and conditional upload control visibility by @Lucifergene in #8066
  • Convert module onboarding rule into a skill by @lucferbux in #7885
  • RHOAIENG-51778, RHOAIENG-58953: vector store error microcopy + Go version docs by @gmenher in #7937
  • chore(deps): expand Dependabot npm config across monorepo by @christianvogt in #7989
  • chore(deps): bump tmp from 0.2.6 to 0.2.7 in /packages/model-registry/upstream/frontend by @dependabot[bot] in #8098
  • chore: update minor/patch deps, bump PF to 6.5 by @christianvogt in https://github.com/opendatahub...
Read more

v3.4.4-odh

Choose a tag to compare

@ppadti ppadti released this 15 Jun 12:12
e62486a

Notable Changes

  • Ongoing features:
    • Validated defaults for tool calling (stage 1)
    • Surface model cold-start and vRAM metrics in Model Catalog
    • NIM wizard + project keys
    • Interactive playground for evaluating AI pipelines without full deployment
    • MaaS Self-Service Subscription Page for Non-Admin Users
    • Prompt Variables - Template Variable Substitution in Playground
    • Global MLflow Workspace for Shared Resources
    • Feature Store & Workbenches Integration Phase 2 — Dashboard Discovery and Visibility
    • Embed perses dashboards
    • EvalHub UI: Side-by-Side Evaluation Run Comparison for AI Artifacts (numeric scores)
    • MLFlow Prompt Template Integration in gen AI studio
    • EvalHub UI: Configurable Quality Thresholds for Benchmarks and Evaluation Collections
    • EvalHub UI: Evaluation Run Configuration Validation and Primary Metric Selection
    • Multimodal Support in Gen AI Studio Playground
  • Miscellaneous bugs and improvements

What's Changed

  • feat(autorag/automl): add missing leaderboard column metadata and descriptions by @nickmazzi in #7396
  • add onyx scrum to jira-triage by @christianvogt in #7376
  • fix: correct gen-ai Tekton pipeline triggers by @christianvogt in #7456
  • fix: Show connected resources for URI-v1 connections by @YuliaKrimerman in #7443
  • fix(autorag): warn when no compatible vector I/O providers found by @jefho-rh in #7450
  • fix(autorag): handle models loading and error states in configure page by @jefho-rh in #7430
  • fix(mlflow): Rename experiments MLflow left nav label to Experiments by @Taj010 in #7461
  • Add Model-registry cypress and unit tests to the CI by @ppadti in #7338
  • feat(RHOAIENG-29749): validate database name in model registry settings by @manaswinidas in #7292
  • MaaS microcopy fixes by @ashley-o0o in #7407
  • fix(gen-ai): return empty slice instead of nil from extractEndpointsFromLLMInferenceService by @NickGagan in #7449
  • fix(gen-ai): reset toolsLoaded before each MCP server tools fetch by @NickGagan in #7451
  • fix(automl,autorag): scale pipeline graph spacing with task label width by @jkyaw in #7377
  • RHOAIENG-58577: Replace rest.CopyConfig with rest.AnonymousClientConfig in gen-ai BFF by @Lucifergene in #7462
  • fix(gen-ai): preserve vector store selection when entering compare mode by @NickGagan in #7447
  • feat(automl,autorag): add reconfigure flow for runs by @nickmazzi in #7352
  • RHOAIENG-31273: Add Cypress mock tests for testing gaps by @rsun19 in #7422
  • feat(automl,autorag): Add error handling and retry for results pages by @nickmazzi in #7409
  • E2E Test for MaaS Subscriptions by @katieperry4 in #7282
  • fix(gen-ai): update AI asset endpoints page copy by @NickGagan in #7452
  • Revoke all my api keys kebab grayed out when disabled by @sridarna in #7460
  • Remove auth annotation when MaaS is selected in Wizard by @Griffin-Sullivan in #7470
  • Fix federated mode theme defaulting to OS preference instead of odh default by @nananosirova in #7345
  • Fix testWorkbenchCreation tests inter-dependencies by @FedeAlonso in #7472
  • test(automl): add AutoML Cypress E2E and mocked tests by @jefho-rh in #7378
  • RHOAIENG-38577: Can't stop a basic workbench on a cluster with custom OIDC provider by @srtanish1992 in #7420
  • fix(manifests): align container resources with K8s best practices and add ephemeral-storage by @lucferbux in #7224
  • feat(manifests): introduce dedicated ServiceAccount for modular architecture BFF modules by @lucferbux in #7238
  • Pass workspace context to external MLflow UI on launch by @nananosirova in #7459
  • overhaul node_modules caching to eliminate excessive cache misses by @christianvogt in #7424
  • docs: add multi-agent workflow guidance documentation by @lucferbux in #7431
  • fixed flagged vulnerabilities in audit by @rsun19 in #7437
  • feat(gen-ai): inline NeMo guardrail config with per-request model discovery by @pnaik1 in #7440
  • Fix: testEnabledISVs cypress e2e test by @FedeAlonso in #7474
  • Upgrade MaaS go version to 1.25.9 by @Griffin-Sullivan in #7448
  • MaaS/Gen-AI Inter-BFF Fixes by @katieperry4 in #7476
  • Add Zahra and Nikhil as reviewers and Conor as approver by @FedeAlonso in #7465
  • fix: prevent Tests summary job from being skipped on failures by @rsun19 in #7489
  • add nim option by @emilys314 in #7453
  • Create rbac evaluation review skill by @rsun19 in #7467
  • update cypress version and address cypress binary caching by @christianvogt in #7492
  • fix(e2e): update default catalog sources ConfigMap name to match operator rename by @YuliaKrimerman in #7374
  • Sync from kubeflow/model-registry 8f14aaf by @ppadti in #7496
  • add guardrails feature flag by @pnaik1 in #7463
  • add approvers and reviewers for cypress, automl and autorag by @tamdavid-git in #7473
  • fix fetch serving settings not loaded when result undefined by @emilys314 in #7490
  • RHOAIENG-59040 Mock tests for prompt management by @danpierce1 in #7329
  • feat(gen-ai): add isEnabled property to Guardrails Enabled tracking e… by @pnaik1 in #7485
  • Fix auth policy view page when model refs do not exist by @katieperry4 in #7432
  • Adding Inter-BFF Communication to MaaS Docs by @katieperry4 in #7495
  • failed a home test by @rsun19 in #7498
  • Revert "failed a home test" by @rsun19 in #7500
  • Add dedicated HWP to testRegistryDeployModel to prevent flaky webhook failures by @nananosirova in #7486
  • Sync from kubeflow/model-registry 8c96833 by @ppadti in #7517
  • chore(deps): bump @babel/plugin-transform-modules-systemjs from 7.28.5 to 7.29.4 in /packages/autorag/frontend by @dependabot[bot] in #7510
  • fix(gen-ai): gate MaaS features on backend service readiness via DSC by @jharan1 in #7446
  • RHOAIENG-58422: Playground: placeholder messages should disappear when real user/assistant messages exist by @jira-autofix[bot] in #7296
  • Release: Tekton pipelines → v3.4.3-odh by @odh-devops-app[bot] in #7528
  • add error code conflict for nemoguardrails by @pnaik1 in #7525
  • CI: Tekton pipeline sync by @odh-devops-app[bot] in #7529
  • feat(gen-ai): add NeMo Guardrails integration to View Code by @NickGagan in #7501
  • testGenAi retryableBefore cleanup deletes assets needed by subsequent tests by @ikeola13 in #7518
  • feat(nim-serving): NIM settings card for project API key management by @mturley in #7436
  • Content updates for Hardware Profile Settings and Kueue integration UI by @Taj010 in #7478
  • fix the pipeline name by @MohammadiIram in #7541
  • Release: Tekton pipelines → v3.4.3-odh by @odh-devops-app[bot] in https://github.com/opendatahub-io/odh-dashboard/pull/...
Read more

v3.4.3-odh

Choose a tag to compare

@manaswinidas manaswinidas released this 04 May 17:34
12e498e

Notable Changes

  • Miscellaneous bugs and improvements

What's Changed

  • Update Tekton files to version v3.0.1-odh by @odh-devops-app[bot] in #5465
  • Upversion dashboard by @ppadti in #5467
  • Update Tekton files to version v3.0.2-odh by @odh-devops-app[bot] in #5738
  • Upversion dashboard by @manaswinidas in #5741
  • Upversion dashboard by @pnaik1 in #5961
  • Red Hat Konflux update mod-arch-maas by @red-hat-konflux[bot] in #5991
  • update tekton push file for odh by @alexxfan in #5992
  • Upversion dashboard by @manaswinidas in #6204
  • Update mod-arch-maas-push.yaml to not use central pipeline for odh release builds by @MohammadiIram in #6222
  • Upversion dashboard by @ppadti in #6619
  • Upversion Dashboard by @manaswinidas in #7070
  • fix(e2e): shorten model registry test fixture prefixes to avoid DNS limit by @mturley in #7188
  • fix: flaky argo workflow and v1 pipeline import tests in pipelines.cy.ts by @caponetto in #7252
  • Feat(RHOAIENG-56731): type guards strengthening by @claudialphonse78 in #7204
  • Fix(RHOAIENG-58437):Text overflow issue for subscription dropdown in create api modal by @claudialphonse78 in #7273
  • feat(mlflow): add BFF status endpoint and DSPA integration by @caponetto in #7119
  • Fix subscription view page when model refs do not exist by @katieperry4 in #7277
  • fix(model-serving): gate AAA checkbox behind genAiStudio feature flag by @Taj010 in #6926
  • Add new general approvers by @manaswinidas in #7278
  • Update oncel exp by @MohammadiIram in #7227
  • feat: Implement BFF inter-communication by @lucferbux in #6404
  • fix(kueue): migrate Kueue API from v1beta1 to v1beta2 by @PR3MM in #7271
  • feat(automl,autorag): support namespace persistence by @MatthewAThompson in #7286
  • Add lint, type-check, and CI workflows for model-registry frontend and BFF by @ppadti in #7122
  • e2e test for vLLM on MaaS by @ashley-o0o in #7004
  • Hide API Key Behind a Toggle by @katieperry4 in #7287
  • chore(automl,autorag): simplify local development set-up when running in federated mode by @chrjones-rh in #7281
  • chore(deps): bump dompurify from 3.3.0 to 3.4.0 in /packages/autorag/frontend by @dependabot[bot] in #7290
  • fix(maas): subscription filter matches displayName and description by @dpanshug in #7270
  • fix(autox): fix 500 error after many duplicate file uploads by @daniduong in #7220
  • Remove api keys empty state and fix status labels by @Griffin-Sullivan in #7300
  • Sort Subscriptions and Policies by Display Name by @katieperry4 in #7294
  • MaaS Microcopy Updates by @ashley-o0o in #7279
  • update jira triage to incorporate observability and perses by @christianvogt in #7311
  • feat: add review agent skill for UI convention compliance by @jenny-s51 in #7064
  • fix(bff): remove unnecessary RequireAccessToService from K8s-only routes (RHOAIENG-55748) by @liday-rh in #7086
  • Sync from kubeflow/model-registry d941024 by @manaswinidas in #7301
  • feat(automl): add One vs Rest view to confusion matrix by @jefho-rh in #7240
  • fix(cypress): fix access mode verification in verifyStorageClassConfig by @srtanish1992 in #7319
  • feat(autorag/automl): add support for stopping and retrying a run by @nickmazzi in #7288
  • Fix(RHOAIENG-58427):Added toolbar with search based on policy name and description by @claudialphonse78 in #7268
  • refactor(autorag): drop openai-go SDK, use typed LlamaStack structs by @jefho-rh in #7283
  • Fix(RHOAIENG-57631):added bff changes for preserving unmanaged field in Policies and subscription by @claudialphonse78 in #7257
  • feat(autox): support selecting files by clicking rows in file explorer by @daniduong in #7114
  • fix(hardware-profiles): add character limits to name and description fields by @kanishka-commits in #6960
  • fix: Use full page navigation for OIDC logout instead of AJAX fetch by @GowthamShanmugam in #7083
  • update inference service check by @sridarna in #7323
  • [Cypress-e2e] add wait for model-catalog pod ready before performance filter test by @ConorOM1 in #7330
  • Fix button order and standardize help icons in Deploy MCP Server modal by @manaswinidas in #7318
  • test: Update Feast tests to avoid feast apply error by @Srihari1192 in #7324
  • fix(automl,autorag): extend dynamic port-forwarding to S3 and LlamaStack paths by @chrjones-rh in #7310
  • home page task assistant by @christianvogt in #7285
  • fix: remove .github exclusion from CodeRabbit path filters by @christianvogt in #7344
  • chore(autorag+automl): improvements - S3 timeout error handling + FileExplorer UX by @GAUNSD in #7219
  • Add agent rule to use PR template when creating pull requests by @christianvogt in #7341
  • chore(deps): bump github.com/moby/spdystream from 0.5.0 to 0.5.1 in /packages/automl/bff by @dependabot[bot] in #7315
  • Evalhub UI updates by @ikeola13 in #7331
  • Sync from kubeflow/model-registry 6e18d5a by @ppadti in #7348
  • update the ux changes by @pnaik1 in #7346
  • fix(model-registry): surface K8s name validation errors on MCP deploy name field by @manaswinidas in #7192
  • feat(autorag/automl): add retry and stop actions to runs table rows by @nickmazzi in #7321
  • chore(deps): bump dompurify from 3.3.3 to 3.4.1 in /packages/automl/frontend by @dependabot[bot] in #7359
  • chore(automl,autorag): refresh embedded pipeline YAMLs from upstream by @chrjones-rh in #7307
  • Fix(RHOAIENG-58940) :Added fix for pause to be disabled for clusterselector ray job by @claudialphonse78 in #7349
  • [Cypress-e2e] add polling step for validated model cards by @ConorOM1 in #7367
  • Add e2e tests for MLflow experiments page by @nananosirova in #7289
  • Disabled actions for deleting subs and policies by @katieperry4 in #7309
  • fix(model-serving): add optional chaining for ServingRuntime spec access by @srtanish1992 in #6855
  • fix(e2e): handle other catalog sources when verifying disable in testSourceEnableDisable by @manaswinidas in #6734
  • [Cypress-e2e] Add model registry custom properties retention test by @ConorOM1 in #7313
  • fix(model-registry): scope in-cluster K8s client to SAR-only interface in MCP deployment auth by @manaswinidas in #7259
  • fix(storage): correct storage percentage for shared and block storage by @DaoDaoNoCode in #6829
  • Add mock tests for mlflow experiments and prompt...
Read more

v3.4.0

Choose a tag to compare

@manaswinidas manaswinidas released this 16 Apr 07:37
04e806c

Notable Changes

  • New features:
    • MaaS API Keys and Quality Hardening
    • MaaS Subscription Model Redesign
    • List Available Subscriptions to "View Endpoint" in AI Asset Endpoint
    • Prompt Management integration in gen AI studio playground
    • AI Asset Endpoints - Add Custom Endpoints Support
    • Surface Existing Vector Stores in Gen AI Studio Playground
    • Inclusion of Ray jobs to existing Training dashboard
    • AutoML UI MVP Experience
    • Eval Hub
    • OCI complaint storage for model registry
    • MCP catalog and deployments (Dev Preview)
  • Ongoing features:
    • Kueue workload workbench scheduling
    • Enable vLLM Runtime Support in MaaS
  • Miscellaneous bugs and improvements

What's Changed

Read more

v3.4.2-odh

Choose a tag to compare

@manaswinidas manaswinidas released this 06 Apr 13:58
096ea11

Notable Changes

  • Ongoing features
    • Kueue workload workbench scheduling
    • MaaS API Keys and Quality Hardening
    • MaaS Subscription Model Redesign
    • List Available Subscriptions to "View Endpoint" in AI Asset Endpoint
    • Prompt Management integration in gen AI studio playground
    • AI Asset Endpoints - Add Custom Endpoints Support
    • Surface Existing Vector Stores in Gen AI Studio Playground
    • Inclusion of Ray jobs to existing Training dashboard
    • AutoML UI MVP Experience
    • OCI complaint storage for model registry
    • MCP catalog and deployments
  • Miscellaneous bugs and improvements

What's Changed

Read more

v3.4.1-odh

Choose a tag to compare

@ppadti ppadti released this 09 Mar 14:11
810a268

Notable Changes

  • v3.4.0EA1 release
  • Ongoing features
    • OCI complaint storage for model registry
    • Perses integration
    • Kueue workload workbench scheduling
    • YAML view in llm-d serving UI
    • AutoRAG UI Experience
    • UX Enhancements - AI Playground
    • Eval Hub
    • MaaS API Keys and Quality Hardening
    • Prompt Management - MLFlow Embedded UI
    • AI Asset Endpoints - External Model Endpoints Support
  • Miscellaneous bugs and improvement

What's Changed

Read more

v3.4.0EA1

Choose a tag to compare

@claudialphonse78 claudialphonse78 released this 23 Feb 12:46
71550c9

Notable Changes

  • New features:

    • Rework gen AI studio playground UI to Align with watsonx.ai Prompt
    • Embed MLflow into pipeline views
    • UI support for PostgreSQL for Model registry
  • Ongoing features:

    • OCI complaint storage for model registry
    • Kueue workload workbench scheduling
    • Perses integration
    • Multi instance chat comparison in playground
    • Gen AI: Guardrails, prompt management
    • Embed MLflow in Experiments page
    • MaaS API Keys and Quality Hardening
  • Miscellaneous bugs and improvements

What's Changed

Read more

v3.3.1-odh

Choose a tag to compare

@manaswinidas manaswinidas released this 09 Feb 12:34
280c63b

Notable changes

  • New features:
    • Rework gen AI studio playground UI to Align with watsonx.ai Prompt
    • Embed MLflow into pipeline views
  • Ongoing features:
    • UI support for PostgreSQL for Model registry
    • OCI complaint storage for model registry
    • Kueue workload workbench scheduling
    • Perses integration
    • Multi instance chat comparison in playground
    • Gen AI: Guardrails, prompt management
    • Embed MLflow in Experiments page
    • MaaS API Keys and Quality Hardening
  • Miscellaneous bugs and improvements

What's Changed

Read more