Litellm ryan apr 4

[ryan-crabbe-berri]

feat(ui): expose Azure Entra ID credential fields in provider form
feat(ui): add per-model rate limits to team edit/info views
fix(ui): use entity key for usage export display

circle ci

https://app.circleci.com/pipelines/github/BerriAI/litellm/73091/workflows/760fb24e-2102-4415-8483-2fa439a718af

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
litellm	Ready	Preview, Comment	Apr 4, 2026 11:41pm

[codspeed-hq]

Merging this PR will not alter performance

✅ 16 untouched benchmarks

---

_{Comparing litellm_ryan-apr-4 (e87f6ca) with main (693ad49)}

[greptile-apps]

Greptile Summary

This PR bundles three independent UI improvements, all well-scoped and clean:

• Azure Entra ID credential fields — tenant_id, client_id, and client_secret are added to the Azure provider entry in provider_create_fields.json, making Service Principal authentication configurable from the provider form UI. A matching unit test verifies field presence, field_type, and required flags.
• Per-model rate limits in TeamInfo — The team edit and info tabs now expose model_tpm_limit / model_rpm_limit stored in team metadata. The edit form uses an Ant Design Form.List with duplicate-model validation; the view tabs render per-model limits from info.metadata?.model_tpm_limit and info.metadata?.model_rpm_limit. The backend correctly moves these top-level fields into metadata via _update_metadata_fields, and the UI strips them from the raw metadata JSON textarea to avoid double-editing.
• Entity key used directly for usage export display — The new resolveEntityDisplay helper uses the entity key as both id and alias (falling back to the key when no team alias map entry exists). This fixes exports for non-team entity types (tags, orgs, customers) where the entity key is already the canonical label.

Issues found:

• One vacuous test in utils.test.ts (see inline comment) — no assertions run under the new logic, but coverage is maintained by the adjacent replacement test.

Confidence Score: 5/5

Safe to merge — no correctness bugs, security issues, or data-loss risks found

All findings are P2 (one vacuous test). Logic changes are correct, well-tested, and consistent with backend contracts.

utils.test.ts has one vacuous test that should ideally be updated, but does not block merge

Important Files Changed

Filename	Overview
ui/litellm-dashboard/src/components/EntityUsageExport/utils.ts	Introduces resolveEntityDisplay helper to use entity key directly for display; all export functions updated consistently
ui/litellm-dashboard/src/components/EntityUsageExport/utils.test.ts	Tests updated for new entity-key display behavior; one pre-existing test is now vacuous
ui/litellm-dashboard/src/components/team/TeamInfo.tsx	Adds per-model rate limits UI editor and read-only display; metadata serialization correctly excludes model_tpm_limit/model_rpm_limit from raw textarea
litellm/proxy/public_endpoints/provider_create_fields.json	Azure provider gains tenant_id, client_id, client_secret fields for Entra ID auth; none are individually required
tests/test_litellm/proxy/public_endpoints/test_public_endpoints.py	New test verifies all three Entra ID fields are present, client_secret is password type, and fields are non-required; uses TestClient only

Sequence Diagram

sequenceDiagram
    participant UI as TeamInfo Edit Tab
    participant API as /team/update
    participant Backend as team_endpoints.py
    participant DB as Prisma DB
    participant UIv as TeamInfo View Tab

    UI->>API: PATCH {model_tpm_limit, model_rpm_limit, metadata}
    API->>Backend: UpdateTeamRequest (model limits as top-level fields)
    Backend->>Backend: _update_metadata_fields(updated_kv)
    Note over Backend: Moves model_tpm_limit/model_rpm_limit into metadata
    Backend->>DB: update team row (metadata contains model limits)
    DB-->>Backend: Updated row
    Backend-->>UI: 200 OK
    UI->>API: GET /team/info (fetchTeamInfo)
    API-->>UIv: team_info.metadata.model_tpm_limit/model_rpm_limit
    UIv->>UIv: Render per-model limits in Rate Limits card

Loading

Comments Outside Diff (1)

1. ui/litellm-dashboard/src/components/EntityUsageExport/utils.test.ts, line 379-386 (link)

   Vacuous test — inner assertion never runs

   After the resolveEntityDisplay refactor, Team ID is always the entity key string (never null, undefined, or "-"), so entryWithoutTeamId is always undefined and the expect inside the if block never executes. The test trivially passes while providing zero coverage.

   The new test at line 388 already covers the correct fallback behavior. Consider replacing this test with one that exercises the new behavior:

   Rule Used: What: Flag any modifications to existing tests and... (source)

_{Reviews (1): Last reviewed commit: "chore: poetry lock" | Re-trigger Greptile}