Joshua Valluru
8d5a675878
fix(mcp): expire temporary OAuth discovery results
2026-09-11 07:02:11 -07:00
Joshua Valluru
eec7c1e7f8
chore: merge current staging for MCP diagnostic compatibility
2026-09-11 07:01:06 -07:00
Joshua Valluru
3fc483d424
fix(mcp): capture bounded error diagnostics without exposing credentials
2026-09-11 06:58:40 -07:00
mateo
8762c664b7
test(registry): cover nemotron reasoning, v4-flash vision and xai/groq deprecation dates
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 13:28:35 +00:00
mateo
4382b86b0f
registry audit 2026-09-11: xai/groq deprecation dates, deepseek-v4-flash vision, perplexity nemotron reasoning
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 13:14:37 +00:00
mateo
598e863510
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_registry_audit_2026_09_10b
2026-09-11 13:04:19 +00:00
mateo
da4052dcd7
test: give the fake pooler a readiness budget that survives a loaded CI worker
...
LiteLLM Rust / rust-lint (push) Has been cancelled
LiteLLM Rust / rust-test (push) Has been cancelled
Terraform Modules / fmt, validate, test (aws) (push) Has been cancelled
Terraform Modules / fmt, validate, test (gcp) (push) Has been cancelled
Terraform Provider / gofmt, vet, build, test (push) Has been cancelled
Terraform Provider / Provider endpoints vs proxy OpenAPI schema (push) Has been cancelled
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 10:18:54 +00:00
joshua-berri
2ca29a9a91
fix(model_armor): gate apply_guardrail raise to non-logging_only and require native guardrails to declare logging_only
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 09:33:57 +00:00
mateo
0199cbcc52
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_deflake_20260905
2026-09-11 09:18:30 +00:00
joshua-berri
39b916f13f
fix(model_armor): decorate apply_guardrail with log_guardrail_information
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 09:07:06 +00:00
joshua-berri
f417d7f739
fix(model_armor): record logging_only verdicts without raising so both scans run
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 08:49:11 +00:00
Chetan Soni
5fa693896f
fix(guardrails): stop 500s on POST /responses when a guardrail rewrites input
2026-09-11 01:21:20 -07:00
joshua-berri
a5cc65f1a9
refactor(custom_guardrail): resolve logging_only output translation in async_logging_hook
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 08:10:35 +00:00
joshua-berri
135ec00b27
feat(model_armor): logging_only mode scans completed streams after delivery
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 08:06:09 +00:00
oliver
560e5da320
fix(proxy): archive the old key only after regenerate validation passes
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 08:02:18 +00:00
oliver
e205be80df
fix(proxy): enforce custom_key_update policy on /key/regenerate
...
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-09-11 07:41:13 +00:00
joshua-berri
9a715df212
Merge pull request #40665 from BerriAI/litellm_fix_openapi_mcp_health_4896
...
fix(mcp): check OpenAPI specifications without native MCP handshakes
2026-09-10 22:02:45 -07:00
Kerry Lu
33ec56ed75
test(e2e): rewrite the Redis timeout test as a locust chaos load test
...
The sequential version sent one request at a time, so a Redis outage never
reached the concurrency where the failed-tracking alert body actually grows.
This drives the proxy with locust against one model group of three mock
deployments, two failing at order 1 and one serving at order 2, so every
request spends its retries on the failing pair and lands on the serving
deployment through the order-based fallback. Two phases, a healthy baseline
and a CLIENT PAUSE WRITE window, and every request must succeed in both.
Latency, RSS and CPU are reported as p50/p90/p99 per phase rather than
asserted on: RSS and CPU come from psutil on the proxy's process tree, since
a multi-worker proxy serves /metrics from the prometheus multiprocess
collector and that drops the process collector's series. Thresholds stay open
until weekly runs give real baselines.
Co-Authored-By: Claude Code <noreply@anthropic.com>
2026-09-10 21:55:23 -07:00
Joshua Valluru
bb9b4c4aef
fix(mcp): render registration refusals without recursion
2026-09-10 21:54:06 -07:00
Joshua Valluru
f04fb748c5
fix(mcp): explain refused OAuth registration and bound discovery retries
2026-09-10 21:02:46 -07:00
mateo-berri
cf1f709681
fix(redis): treat asyncio.TimeoutError as a timeout on Python 3.10
2026-09-10 20:59:33 -07:00
mateo-berri
1fe6984ec0
refactor(redis): walk the exception cause chain iteratively
...
The recursion detector flags any unignored recursive function, so the
timeout classification now walks the explicit cause chain with a
bounded generator instead of calling itself
2026-09-10 20:35:25 -07:00
Joshua Valluru
ca03c889c9
fix(mcp): avoid caching cancelled OpenAPI health probes
2026-09-10 20:26:53 -07:00
mateo-berri
fea4970159
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_redis_pool_timeout_counts_as_timeout
2026-09-10 20:23:25 -07:00
joshua-berri
acb9086f29
Merge pull request #40664 from BerriAI/litellm_fix_mcp_vscode_dcr_7449
...
fix(mcp): accept VS Code OAuth registration callbacks
2026-09-10 20:17:46 -07:00
mateo-berri
54e247998e
fix(redis): count pool wait timeouts as breaker timeouts
...
redis-py's blocking pool reports a saturated pool as ConnectionError chained
from asyncio.TimeoutError. The circuit breaker classified that as a hard
connectivity failure and opened at once while Redis was healthy. Follow the
explicit cause chain so it counts as a timeout and stays behind the
timeout_min_duration gate
2026-09-10 20:14:50 -07:00
Joshua Valluru
da1dfcdb24
refactor(mcp): reuse the shared HTTP handler for bounded probes
2026-09-10 19:59:07 -07:00
Mateo Wang
ff4b558243
Merge pull request #40624 from BerriAI/litellm_redis_breaker_open_silent_miss
...
fix(caching): keep an open Redis circuit breaker open and quiet on the sync read and spend counter paths
2026-09-10 19:55:22 -07:00
mateo-berri
77a0053a20
test(e2e): require the same workers at both memory checkpoints
...
Each checkpoint now samples until no new worker has answered for the settle
window, and the growth assertion refuses a worker set that changed between the
warm and after checkpoints instead of comparing only the intersection, so a
leaking worker reached by one checkpoint alone cannot drop out of the gate
2026-09-10 19:47:11 -07:00
Joshua Valluru
576c1bc5d6
fix(mcp): bound and coalesce OpenAPI health probes
2026-09-10 19:45:44 -07:00
mateo-berri
25ed0abfc9
chore(ui): regenerate schema.d.ts after merging the base
2026-09-10 19:42:27 -07:00
mateo-berri
55965cf7fb
chore(ui): regenerate schema.d.ts after merging staging
2026-09-10 19:41:52 -07:00
mateo-berri
fdd423128e
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_redis_breaker_open_silent_miss
2026-09-10 19:40:52 -07:00
mateo-berri
f384acb840
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_e2e_memory_regression_failing_requests
2026-09-10 19:40:11 -07:00
Hayden Moulds
22b4656cb5
chore: synchronize generated proxy API types
2026-09-11 12:40:02 +10:00
Joshua Valluru
5735587133
chore(ui): sync API descriptions with the current default branch
2026-09-10 19:30:10 -07:00
Joshua Valluru
86c5cd736b
chore: sync API descriptions from the updated base branch
2026-09-10 19:29:19 -07:00
Joshua Valluru
fc95d22367
fix(mcp): accept VS Code OAuth registration callbacks
2026-09-10 19:28:40 -07:00
mateo-berri
2fc520329f
fix(router): keep the budget push off the request callback path
...
The provider budget push runs inside the request success callback, so
awaiting the Redis pipeline there made every request wait for the round
trip. Hand it back to a task whose failure is logged through the breaker
aware logger, so an open breaker stays a debug line and a real Redis error
is one error line instead of an unretrieved task traceback
2026-09-10 19:26:28 -07:00
mateo-berri
eca7bb11ce
fix(proxy): read RSS from /proc when psutil is missing so the release image reports memory
...
The release image installs only the proxy extras, and psutil is a locust and mirakuru dev
dependency, so /debug/memory/summary answered with an error and no ram_usage_mb on the e2e
gate. Fall back to /proc/self/statm and /proc/meminfo on Linux when psutil cannot be imported
2026-09-10 19:25:55 -07:00
Hayden Moulds
5b9244105c
test(proxy): use deployment listing metadata in alias coverage
2026-09-11 12:23:48 +10:00
Hayden Moulds
2f8f3d3d21
fix(proxy): build public team alias response immutably
2026-09-11 12:22:07 +10:00
Hayden Moulds
e18d766f53
test(proxy): consolidate team alias metadata coverage
2026-09-11 12:22:07 +10:00
Hayden Moulds
e664500003
test(proxy): cover team alias retrieve metadata
2026-09-11 12:21:57 +10:00
Hayden Moulds
49809a814b
fix(proxy): preserve metadata for public team aliases
2026-09-11 12:21:57 +10:00
yuneng-jiang
362033cb2a
Merge pull request #40658 from BerriAI/litellm_internal_staging
...
Terraform Provider / gofmt, vet, build, test (push) Has been cancelled
Terraform Provider / Provider endpoints vs proxy OpenAPI schema (push) Has been cancelled
Unit Tests: Documentation Validation / documentation (push) Has been cancelled
GitHub Actions Security Analysis / zizmor (push) Has been cancelled
Unit Tests / caching-local (push) Has been cancelled
Unit Tests / core-utils (push) Has been cancelled
Unit Tests / enterprise-package (push) Has been cancelled
Unit Tests / enterprise-routing (push) Has been cancelled
Unit Tests / integrations (push) Has been cancelled
Unit Tests / All Other Providers (push) Has been cancelled
Unit Tests / Vertex AI (push) Has been cancelled
Unit Tests / misc (push) Has been cancelled
Unit Tests / proxy-auth (push) Has been cancelled
Unit Tests / proxy-endpoints (push) Has been cancelled
Unit Tests / proxy-extras (push) Has been cancelled
Unit Tests / proxy-infra (push) Has been cancelled
Unit Tests / proxy-server (push) Has been cancelled
Unit Tests / responses-caching-types (push) Has been cancelled
Unit Tests: Proxy DB Operations / auth-checks (push) Has been cancelled
Unit Tests: Proxy DB Operations / budgets (push) Has been cancelled
Unit Tests: Proxy DB Operations / custom-logging (push) Has been cancelled
Unit Tests: Proxy DB Operations / db-and-spend (push) Has been cancelled
Unit Tests: Proxy DB Operations / endpoints-and-responses (push) Has been cancelled
Unit Tests: Proxy DB Operations / guardrails-hooks (push) Has been cancelled
Unit Tests: Proxy DB Operations / jwt-and-keys (push) Has been cancelled
Unit Tests: Proxy DB Operations / key-generation (push) Has been cancelled
Unit Tests: Proxy DB Operations / logging-misc (push) Has been cancelled
Unit Tests: Proxy DB Operations / proxy-runtime (push) Has been cancelled
Unit Tests: Proxy DB Operations / proxy-server-core (push) Has been cancelled
Unit Tests: Proxy DB Operations / proxy-utils (push) Has been cancelled
chore(ci): promote internal staging to main
2026-09-10 19:19:44 -07:00
mateo-berri
0c84f656b5
Merge remote-tracking branch 'origin/litellm_internal_staging' into litellm_ocr_custom_pricing
LiteLLM Rust / rust-lint (push) Has been cancelled
LiteLLM Rust / rust-test (push) Has been cancelled
Terraform Modules / fmt, validate, test (aws) (push) Has been cancelled
Terraform Modules / fmt, validate, test (gcp) (push) Has been cancelled
Terraform Provider / gofmt, vet, build, test (push) Has been cancelled
Terraform Provider / Provider endpoints vs proxy OpenAPI schema (push) Has been cancelled
2026-09-10 19:18:43 -07:00
tin-berri
7419a536ad
fix(auto-router): omit Claude Code system text from classifier ( #40655 )
...
Co-authored-by: Claude Code <noreply@anthropic.com>
2026-09-10 19:18:25 -07:00
Joshua Valluru
8c82c325ac
fix(mcp): check OpenAPI specifications without native MCP handshakes
2026-09-10 19:18:19 -07:00
mateo-berri
0ffe6512de
fix(router): keep the routing and budget sync loops quiet while the Redis breaker is open
2026-09-10 19:15:40 -07:00
