Commit graph

16833 commits

Author SHA1 Message Date
yassin
ce0301c23f refactor(proxy): drop redundant docstrings from upload allowlist helpers and tests
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 19:00:49 +00:00
kerry
250ff03a04 feat(model_info): scope fill_missing rules to azure, bedrock and vertex hosts
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:58:53 +00:00
kerry
812bbee0b3 fix(model_info): scope fill_missing backfill to the rule's providers
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:58:53 +00:00
Devin AI
fb2057fde7 refactor(model_info): rename backfill_exact_entries to fill_missing_fields
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:58:53 +00:00
Devin AI
42c708670d fix(model_info): guard backfill by mode, drop provider key, tighten claude major regex
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:58:53 +00:00
Devin AI
cee7215b24 feat(model_info): opt-in field-level backfill from fallback generalization rules
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:58:53 +00:00
yucheng-berri
cae4a65545
Merge pull request #40562 from BerriAI/litellm_otel_openinference_message_cap
fix(otel): cap per-index OpenInference message attributes span-wide
2026-09-14 11:55:37 -07:00
yucheng
80725addb4 Merge remote-tracking branch 'origin/main' into litellm_lit6314_guardrail_metadata_transfer 2026-09-14 18:54:24 +00:00
yassin
82ef6ea6ab feat(proxy): add general_settings.allowed_file_extensions for /v1/files
Opt-in allowlist for upload filename extensions, checked before the existing blocked_file_extensions blocklist and mapped through the same upload validation failure path. None keeps today's behaviour, [] rejects every upload, matching is case-insensitive on both sides, and a filename with no extension is rejected when the allowlist is set.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:53:29 +00:00
Yujong Lee
d4869ba710 refactor: print rust and python traces independently 2026-09-14 11:48:25 -07:00
Yujong Lee
eb2ffaae79 fix(harness): rebuild bridge for trace parity 2026-09-14 11:48:25 -07:00
Yujong Lee
561533c596 wip 2026-09-14 11:48:24 -07:00
Yujong Lee
ad5b87eb91 fix: capture reused worker threads in Python traces 2026-09-14 11:47:30 -07:00
Mateo Wang
cab1e113f7
Merge pull request #40976 from BerriAI/litellm_azure_gpt_chat_latest_pricing
feat(pricing): add azure gpt-chat-latest rates and drop retired friendliai llama-3.1 entries
2026-09-14 11:21:20 -07:00
Yassin Kortam
08a78a3982
Merge pull request #41054 from BerriAI/litellm_gate_correlation_contextvar_stamp
perf(logging): skip correlation contextvar stamping when request_correlation_in_logs is off
2026-09-14 11:14:44 -07:00
Shivi Jain
f4e21430c0 test(auth): freeze prefetch cache clock so the org getter cannot miss on a slow runner
Prefetch still writes org entries with the 5s getter TTL. This test only asserts the SQL join, and wall-clock expiry on CI turned that into a MagicMock await TypeError.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-09-14 18:13:49 +00:00
Shivi Jain
75b16df6fa fix(auth): keep prefetched org entries on the 5s getter TTL
Organization mutations do not evict those cache keys, so stretching prefetch to the management TTL would leave stale org grants in L1.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-09-14 18:13:49 +00:00
Shivi Jain
abaa2f8b81 fix(auth): put TQ008 suppressions on the patch call lines
The test-quality gate attributes the comment to the `patch(` line, so reasons on the closing paren did not count and lint failed after format started passing.

Co-authored-by: Cursor <cursoragent@cursor.com>
2026-09-14 18:13:49 +00:00
Shivi Jain
53ba8b9866 fix(auth): load team membership once per request and skip prisma on an L1 hit
common_checks was querying get_team_membership twice, and DualCache awaited Redis SET on the auth path, so LRU eviction plus a hung Redis write showed up as two postgres spans
2026-09-14 18:13:49 +00:00
Shivi Jain
70ddc7e492 fix(auth): load team membership once per request and skip prisma on an L1 hit
common_checks was querying get_team_membership twice, and DualCache awaited Redis SET on the auth path, so LRU eviction plus a hung Redis write showed up as two postgres spans
2026-09-14 18:13:49 +00:00
Shivi Jain
5457f48290 fix(auth): load team membership once per request and skip prisma on an L1 hit
common_checks was querying get_team_membership twice, and DualCache awaited Redis SET on the auth path, so LRU eviction plus a hung Redis write showed up as two postgres spans
2026-09-14 18:13:49 +00:00
joshua-berri
eceb67b69a test(model_armor): inject HTTP handler and token provider
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 18:03:22 +00:00
yucheng
4be84a859d Merge remote-tracking branch 'origin/main' into litellm_lit6314_guardrail_metadata_transfer 2026-09-14 18:01:01 +00:00
Yuneng Jiang
1d49e0ee7e
test: preserve immutable integration observations and cleanup outcomes 2026-09-14 10:58:36 -07:00
Yuneng Jiang
4ee73c49af
test: collect pricing observations immutably 2026-09-14 10:58:24 -07:00
Mateo Wang
6e9e08475a
Merge pull request #40627 from BerriAI/litellm_fix_realtime_cached_audio_cost
fix(cost): bill cached realtime audio tokens at the audio cache-read rate
2026-09-14 10:50:00 -07:00
yucheng
34e702c571 test(otel): build the clipped-blob conversation without mutating the payload
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 17:46:18 +00:00
yucheng
b4e3d7c533 chore: merge main into litellm_lit6314_guardrail_metadata_transfer
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 17:33:04 +00:00
yassin
83d16a4690 fix(proxy): read LITELLM_DISABLE_ACCESS_LOG_PATHS per record and match before redaction
Values loaded after import via proxy config environment_variables or dotenv
were ignored, and a long query string was truncated by the redaction filter
before the path filter could match it. Tests now go through the production
registration on the uvicorn.access logger instead of a hand-built filter.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 17:27:43 +00:00
Devin AI
66ce1eea97 test(proxy): cover non-string access log paths
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 16:54:38 +00:00
yassin
4b39344d13 fix(proxy): only treat a missing user as no memberships on team list, surface db errors
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 16:49:54 +00:00
Devin AI
2268bbaf5e feat(proxy): honor LITELLM_DISABLE_ACCESS_LOG_PATHS to drop noisy uvicorn access log lines
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 16:39:54 +00:00
yassin
dc30be006b refactor(proxy): read org admin's own team ids via get_user_object(check_db_only=True)
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 16:27:31 +00:00
Yuneng Jiang
96fdf5d6f4
test: isolate management integration coverage from E2E registry 2026-09-14 09:23:00 -07:00
Yuneng Jiang
48ccdba129
test: decouple integration contracts from E2E coverage registry 2026-09-14 09:22:13 -07:00
yassin
d643909789 fix(proxy): read org admin's own team ids from db and assert v2 list results
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 16:03:23 +00:00
yassin
8d9edfc03c fix(proxy): keep /team/list self query for org admins membership-only across orgs
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 15:40:41 +00:00
yassin
58c3ccf668 fix(proxy): keep org admins' own team memberships in other orgs visible on team list
An org admin listing their own teams on GET /team/list and GET /v2/team/list
only saw teams in the orgs they administer. Teams they belong to in other
orgs were dropped because the org scope and the membership scope were ANDed.
A self query now unions the two, while a query for another user keeps the
org boundary intersection.

Resolves LIT-3723

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 15:27:06 +00:00
Yassin Kortam
9e1ed40db3
Merge pull request #40927 from BerriAI/litellm_per_issuer_virtual_key_claim_field
feat(jwt): allow virtual_key_claim_field per issuer
2026-09-14 08:00:25 -07:00
Yuneng Jiang
cfda6dea3e
test: cover persisted updates and warmed authorization policies 2026-09-14 04:20:06 -07:00
Yuneng Jiang
92e0b72e2d
test: add CircleCI integration contract foundation 2026-09-14 03:30:52 -07:00
yassin
15ff8d18e7 test(proxy): cover the CLI single-model branch of /model/info JSON serialization
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 10:07:10 +00:00
yassin
b86de179dc fix(utils): stop wrapper_async submitting the sync success handler twice
_client_async_logging_helper re-submitted logging_obj.success_handler to the
executor after _dispatch_success_logging had already done so, running the same
success pipeline twice per async request and racing on shared logging state.

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 10:02:51 +00:00
yassin
93e6770d68 perf(proxy): serialize /model/info listing once with orjson
FastAPI walked the multi-megabyte /model/info payload through jsonable_encoder
before json.dumps on every request. Return a prebuilt orjson Response instead,
keeping jsonable_encoder as the fallback for datetimes and other non-native values

Resolves LIT-5724

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 09:37:00 +00:00
Devin AI
a70f9a17cc perf(logging): skip correlation contextvar stamping when request_correlation_in_logs is off
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-14 08:54:34 +00:00
Oliver Jensen
f9da8a19b6
test(models): stop the password serialization test matching field-name substrings 2026-09-14 09:59:45 +02:00
Oliver Jensen
40118bd158
test(auth): annotate the session-minting patch for the test-quality gate 2026-09-14 09:59:45 +02:00
Oliver Jensen
e77d11d8d7
refactor(auth): type the breach-screen DB dicts and flatten the session-key kwargs
Annotate screen_login_password_for_breach's update/where dicts with
prisma input TypedDicts and replace authenticate_user's conditional
dict splat with plain keyword arguments, clearing the LIT002 lines
this branch added in login_utils.py. No behavior change: an unflagged
login now passes allowed_routes=None and metadata={} explicitly, which
are the parameter defaults
2026-09-14 09:59:45 +02:00
Oliver Jensen
4f2836bc60
feat(auth): screen the login password inline and restrict the session on a fresh breach hit
A breach found during a login previously only flagged the account for the
NEXT login, handing out one free unrestricted 24h session. The HIBP screen
is now awaited before the session key is minted (worst case one 5s window
per user per 24h, fail-open unchanged), so a fresh hit restricts the
current session and the dashboard routes straight to change-password.

Also repairs two casualties of merge f5e47974db that the layout tests
caught: the lost usePathname import and a call to migratedHref, which
staging renamed to uiHref.
2026-09-14 09:59:45 +02:00
Oliver Jensen
671d032b20
feat(auth): force password reset for breached or admin-set passwords 2026-09-14 09:59:45 +02:00