dev #1

ivanwu · 2026-05-22T09:39:52Z

ivanwu commented

2026-05-22 09:39:52 +00:00

v1.2.1 with user auth, scale up/down, streaming logs and helm upgrade with values.yaml. Bugs still: quato limit bwtween cluster, deplicated key when creating whatever cluster or namespace and so on.

ivanwu added 15 commits 2026-05-22 09:39:52 +00:00

refactor: full-stack restructure with multi-tenancy, workspace management, and K8s diagnostics 7f238a3168

- Add Workspace domain (entity, repository, service, handler, DTO)
- Add multi-tenant K8s client with tenant binding and quota management
- Add K8s diagnostics client (instance diagnostics)
- Add authorization middleware (authz package)
- Restructure frontend to feature-based architecture (features/)
- Add User Management page in configuration
- Add AccessDenied page and route guards
- Refactor shared components (form inputs, layout, UI)
- Update Tailwind config for new design system
- Add comprehensive documentation (docs/, tasks/, plans)
- Improve cluster service with better kubeconfig handling
- Add tests for crypto, config, helm client, tenant binding

feat: fix YAML field conversion, admin namespace, streaming logs, and vllm-serve deploy 7d9545f827

- Fix Axios keysToSnake converting user values map keys (gpuMem -> gpu_mem)
  - Add skipRecurseKeys to keysToSnake for values/valuesYaml fields
  - Add values_yaml alt json tag and Normalize() in DTOs
  - Check both camelCase/snake_case in enforceNamespaceValues
  - Read both tailLines/tail_lines query param for diagnostics
- Admin users can freely choose namespace in LaunchModal (free-text input)
  - Block only kube-system/kube-public/kube-node-lease for admin
  - Regular users keep existing namespace restrictions
- Add SSE streaming pod logs endpoint (backend + frontend)
  - New PodLogStreamer interface and K8s Follow:true implementation
  - SSE handler with text/event-stream output
  - Frontend DiagnosticsModal: Stream button, auto-scroll, live indicator
- Remove per-card Refresh button from InstanceCard (redundant with page refresh)
- Deploy bge-m3 on vllm-serve 0.6.0 (gpuMem=10000, status=deployed)

fix: remove per-card Refresh button in InstanceCard, consolidate to page-level refresh 87eaaa564b

- Remove onRefresh prop from InstanceCard component
- Remove handleRefresh callback from InstancesManagementPage
- Remove onRefresh passing from both filtered and unfiltered instance lists
- Keep global page refresh button in PageHeader

feat: scale instances, --reuse-values, values diff, UI redesign, hover animations 28ecb2e636

Backend (Phase 1):
- Add ScaleInstance endpoint (POST /clusters/{id}/instances/{id}/scale)
- Add GetInstanceValuesDiff endpoint (GET .../values-diff)
- Enable ReuseValues=true in Helm Upgrade for --reuse-values behavior
- Add GetValues/GetChartDefaultValues to HelmClient interface
- Add ScaleInstanceRequest/Response and InstanceValuesDiffResponse DTOs

Frontend (Phase 2):
- InstanceCard: +/- scale buttons with loading spinner
- ModifyModal: values diff view (current vs defaults), Use Defaults button
- ArtifactBrowserPage: collapsible sidebar, compact tag grid, search filter
- TagCard: "LATEST" badge, compact layout, responsive design
- InstanceCard: compact 3-column layout, fewer scrolls needed
- InstancesManagementPage: 3-column grid, compact view
- Global hover-lift and hover-glow CSS utilities
- SidebarNav: subtle hover transition on links

fix: UI redesign — horizontal instance rows, proper scaling, readable tag cards 49b92e66c3

- Backend: add replicas field to InstanceResponse (extracted from values.replicaCount)
- InstanceCard: complete redesign as horizontal row layout
  - Status bar | Name+Chart | Replicas +/- | Action buttons
  - Scale controls show for deployed AND failed statuses (scale to 0)
  - Fix replicas display using new instance.replicas backend field
- InstancesManagementPage: vertical row list + onScale callback to update state
- TagCard: restore proper padding (p-4), min-width, readable button sizes
- ArtifactBrowserPage: reduce grid density (sm:1 md:2 lg:3)
- ModifyModal: simplify to YAML-only editing with current values pre-populated
  - Remove schema-based form generator
  - Keep values-diff as collapsible reference panel

fix: direct K8s scaling, replicas from K8s API, button labels, modify fetch 4441f58299

- Add ScaleClient using K8s API (like kubectl scale deploy --replicas=N)
  - ScaleDeployment: patch Deployment.Spec.Replicas directly
  - GetDeploymentReplicas: query actual K8s deployment replicas
  - Search by labels then fallback to deployment name match
- Wire ScaleClient to InstanceService via SetScaleClient in main.go
- ModifyModal: fetch full instance detail on open (list excludes values)
- InstanceCard: add text labels to action buttons (Entries/Diag/Modify/Delete)
  - Text visible on sm+ screens, icon-only on xs

fix: scale replicas in response, YAML lineWidth, delta values, modified keys 96d42ee3e1

- Scale API now returns actual replicas in instance response
- ModifyModal: fix YAML stringify line breaking (lineWidth: 0)
- ModifyModal: show modified keys summary above YAML editor
- ModifyModal: only send delta (user-modified) values to server
- Add diffObjects helper for deep object comparison

fix: real K8s replicas in list API, full Helm values in modify YAML editor b88fe24aab

- Add Replicas field to entity.Instance
- Add EnrichReplicas to InstanceService (batch K8s deployment query)
- convertInstanceResponse uses instance.Replicas instead of hardcoded 0
- ModifyModal: load full Helm values from values-diff API (Current deployed)
- Remove stale loadValuesDiff, use single useEffect for all data loading
- Fix YAML lineWidth:0 for no line wrapping

chore: cleanup old docs, add regression report and unresolved bugs 8f90cf0f0d

- Remove stale test docs (test2-*, bugs-user-*, bug-report, session-notes)
- Add UNRESOLVED-BUGS.md and regression-full-report.md
- Update todo.md

fix: scale replicas in response, K8s metrics client, quota precheck, auth tests 33ddaf97db

- Add GetMetrics method to MetricsClient interface and implement cluster metrics API
- Add QuotaPrecheck service for validating resource quotas before deployment
- Add auth DTO with role/permission models and auth handler tests
- Add instance diagnostics: mounted NFS volumes, labels, annotations in pod diagnostics
- Update workspace handler with GetWorkspace endpoint and shared-user list
- Fix monitoring handler to use correct service method name
- Add tail_lines fallback in instance handler for snake_case query params
- Update nginx config for SSE log streaming support (no buffering)
- Add comprehensive test coverage: auth_service_test, auth_handler_test,
  auth_dto_test, metrics_client_test, quota_precheck_test
- Update error messages for quota validation and instance operations
- ModifyModal: fix YAML lineWidth:0, modified keys summary, delta-only submit
- InstanceCard: correctly disable scale-minus when replicas <= 0
- SidebarLayout: add hover transition for sidebar items
- Update todo.md and lessons.md with latest fixes

fix: auto-enable cluster bootstrap, add init-db.sql to postgres volumes 0144e9cab7

- Auto-enable cluster seeding when BOOTSTRAP_CLUSTERS is set
  (no longer requires separate BOOTSTRAP_ENABLE_CLUSTERS=true)
- Add BOOTSTRAP_ENABLE_CLUSTERS to README .env template
- Mount init-db.sql in postgres service volumes

feat: first-time setup flow — no .env required for deployment 0094519f52

- Add GET /auth/status endpoint (returns needsSetup when no admin exists)
- Add POST /auth/setup endpoint (public first-admin registration)
- Add IsAdminExists + SetupInitialAdmin methods to AuthService
- Frontend: detect needsSetup on load, show setup page with admin registration
- Frontend: fall back to login page when setup is already complete
- Docker compose: env_file already optional (required: false), no changes needed
- Bootstrap: auto-detect BOOTSTRAP_CLUSTERS without separate enable flag

docs: update README for .env-optional deployment and first-time setup flow 7d297a2b1a

- .env is now optional — deployment works with docker compose up --build directly
- First user registers as admin through initial setup page
- Add verification steps for auth/status and auth/setup endpoints

refactor: simplify setup flow — eliminate redundant DB calls and login round-trips e73b3147ed

- Add AdminExists() to UserRepository (EXISTS query, not full table scan)
- SetupInitialAdmin returns tokens directly (skip separate Login call)
- Add SetupRequest DTO to auth_dto.go (replace inline struct)
- Extract defaultEmail() helper (removes duplicated email logic)
- AuthPage uses setup tokens directly (skip redundant apiLogin call)
- Use customAxiosInstance for auth API calls (consistent with codebase)

fix: clear tsbuildinfo cache in frontend-build to prevent stale builds 17acb7d018

- Add rm -rf node_modules/.tmp before npm run build
- Without this, tsc -b reads stale .tsbuildinfo from Docker volume
  and skips recompiling changed files, causing old JS to be served

ivanwu merged commit 66c08e8bc6 into main

2026-05-22 09:41:12 +00:00

ivanwu referenced this issue from a commit

2026-05-22 09:41:13 +00:00

Merge pull request 'dev' (#1) from dev into main

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: OCDP/ocdp-go#1