Commit Graph
75 Commits
Author SHA1 Message Date
Chang Su a2d4f58a96 fix(http): use 504 Gateway Timeout for upstream timeouts (#16320) 2026-01-02 16:01:44 -08:00
Chang Su f66b091699 refactor(http): improve pd_types.rs documentation and style (#16319) 2026-01-02 16:01:11 -08:00
Simo LinandChang Su f39382c6ca [model-gateway] Fix duplicate classify prefix in response ID (#16101)
Co-authored-by: Chang Su <chang.s.su@oracle.com>
2025-12-29 11:07:21 -08:00
Simo LinandChang Su 2ff289e2dc [model-gateway] Generate UUID-based request IDs for embedding/classify (#16100)
Co-authored-by: Chang Su <chang.s.su@oracle.com>
2025-12-29 10:49:58 -08:00
Simo LinandChang Su b2a3f0551a [model-gateway] Wire classify pipeline to gRPC router (#16098)
Co-authored-by: Chang Su <chang.s.su@oracle.com>
2025-12-29 10:44:45 -08:00
Simo LinandChang Su f44c4b3715 [model-gateway] Add classify pipeline stages and protocol types (#16094)
Co-authored-by: Chang Su <chang.s.su@oracle.com>
2025-12-29 10:23:36 -08:00
Simo LinandChang Su 162d1cf9be [model-gateway] Add classification model support infrastructure (#16061)
Co-authored-by: Chang Su <chang.s.su@oracle.com>
2025-12-29 08:34:05 -08:00
Simo Lin c31f62722c [model-gateway] fix tokenizer to match transformers special token handling (#16087) 2025-12-29 08:13:03 -08:00
Simo Lin f13949e5a9 [model-gateway] perf: optimize observability logging for minimal CPU/memory overhead (#16039) 2025-12-28 21:20:46 -08:00
Simo Lin 3645ed0f73 [model-gateway] Add PrefixHash load balancing policy for KV cache-aware routing (#15935) 2025-12-27 05:58:05 -05:00
Simo Lin 171912a9e3 [model-gateway] Add consistent hashing for ManualPolicy routing (#15907) 2025-12-26 22:53:03 -08:00
Simo Lin caa95c7eb4 [model-gateway] update ManualPolicy with header-based routing (#15847) 2025-12-25 13:04:43 -08:00
fzyzcjy 6ef543f99d Use X-SMG-Routing-Key header instead of json body and add tests (#15826) 2025-12-25 09:58:14 -08:00
fzyzcjy ccf2330b28 Tiny fix missing record_router_upstream_response (#15811) 2025-12-25 15:20:43 +08:00
fzyzcjy 45adad37d0 Add manual routing policy for router (#15586) 2025-12-25 11:27:36 +08:00
fzyzcjy 1ba897f330 Tiny refactor select_workers API for future passing more information (#15596) 2025-12-25 10:31:10 +08:00
Simo Lin 643aeefe6d [model-gateway] Fix logging module name, parse endpoint context, and tokenizer factory (#15782) 2025-12-24 10:54:48 -08:00
Simo Lin 2f7c629235 [model-gateway] Fix IGW routing and optimize RouterManager (#15741) 2025-12-24 11:28:53 -05:00
Arthur Cheng f65fa04748 [model-gateway]Enable IGW mode with gRPC router and auto enable IGW when service discovery is turned on (#15459) 2025-12-24 00:27:58 -08:00
Simo Lin 846953d9f1 [model-gateway] Add tokenize/detokenize HTTP endpoints and tokenizer management (#15702) 2025-12-23 17:32:07 -08:00
Simo Lin aa6ac96635 [model-gateway] Fix tokenizer caching and improve error handling (#15695) 2025-12-23 10:33:38 -08:00
ratish 5f3a47d8a7 [model-gateway]: add gRPC router embeddings endpoint implementation (#15273) 2025-12-23 09:12:22 -08:00
Praneth Paruchuri 80ae2229d3 [model-gateway] Optimize router selection with lock-free snapshots (#15672) 2025-12-23 09:09:30 -08:00
Arthur Cheng dd620987d1 [model-gateway] Replace tokenizer with tokenizer registry for dynamic tokenizer loading in gRPC router (#12968) 2025-12-23 07:58:47 -08:00
fzyzcjy b736a1525a Tiny add back missing router per attempt response metric (#15621) 2025-12-22 09:59:10 -08:00
fzyzcjy 6984837d82 Fix router gRPC mode launch error caused by async loading (#15368) 2025-12-22 09:39:01 -08:00
Simo Lin 2142881b7e [model-gateway] return 503 when all workers are circuit-broken (#15611) 2025-12-22 04:06:09 -10:00
Simo Lin 1167867e8d [model-gateway] add retry support to OpenAI router chat endpoint (#15589) 2025-12-21 20:28:51 -08:00
Simo Lin 122c250336 [model-gateway] add retry and circuit breaker support to gRPC routers (#15585) 2025-12-21 17:12:36 -08:00
Simo Lin e1dcd0df3e [model-gateway] minor code clean up (#15578) 2025-12-21 11:00:32 -10:00
Haoran Wang ef001fb811 [model-gateway] /parse/easoning and parse/function_call for sgl-model-gateway (#15568) 2025-12-21 07:39:24 -10:00
Simo Lin 5529ab5895 [model-gateway] Implement RAII load guard with response body attachment (#15507) 2025-12-19 19:14:52 -08:00
kun-llfl 3c116d5e5a [router] bugfix: cache_aware in grpc inbalance forward (#15473)
Signed-off-by: Kun(llfl) <i@imux.top>
2025-12-19 16:25:46 -08:00
Simo Lin ba72e759ca [model-gateway] Improve logging across core modules (#15497) 2025-12-19 13:35:28 -08:00
Simo Lin 53e151945a [model-gateway] optimize worker registry and reduce lock contention in grpc client fetch (#15336) 2025-12-17 08:49:10 -10:00
Simo Lin d747147a26 [model-gateway] reduce cpu overhead (#15316) 2025-12-17 09:42:01 -08:00
Simo Lin 3518b33178 [model-gateway] Remove legacy RouterMetrics and Rename SmgMetrics to Metrics and smg_labels to metrics_labels (#15160) 2025-12-15 04:57:42 -08:00
Simo Lin 7b9156c773 [model-gateway] add mcp and discovery metrics (#15156) 2025-12-14 22:54:53 -08:00
Simo Lin bd9c3a47d6 [model-gateway] Add streaming metrics for harmony gRPC router (#15147) 2025-12-14 20:54:43 -08:00
Simo Lin fb96669ff9 [model-gateway] Add Layer 3 worker metrics (smg_worker_*) (#15130) 2025-12-14 20:25:46 -08:00
fzyzcjy 62b3fdae43 Fix cache aware wrong routing caused by incorrect load tracking (#15101) 2025-12-15 10:03:54 +08:00
Simo Lin 0612175ce1 [model-gateway] add streaming metrics (TTFT, TPOT, tokens, duration) for gRPC router (#15125) 2025-12-14 14:39:10 -08:00
Simo Lin b11af135eb [model-gateway] feat(metrics): implement Layer 2 router metrics (smg_router_*) (#15124) 2025-12-14 10:34:12 -08:00
Simo Lin 74ea45cc4b [model-gateway] optimize metric labels to avoid unnecessary allocations (#15095) 2025-12-13 21:10:45 -08:00
Simo Lin e048ee90fc [model-gateway] Simplify error response creation (#15079) 2025-12-13 10:28:24 -08:00
fzyzcjy 993278b488 Fix double decrease load (#15060) 2025-12-13 22:12:35 +08:00
fzyzcjy 9e9d910744 Fix load metric not updated when using guard (#15059) 2025-12-13 22:11:52 +08:00
fzyzcjy bf17e769fe Add sgl_router_attempt_http_responses_total for single attempt information (#15037) 2025-12-13 13:29:18 +08:00
fzyzcjy 9a5d6a84ab Add error code in prometheus metrics and add X-SMG-Error-Code header (#15036) 2025-12-13 13:28:32 +08:00
fzyzcjy 31c23e5fe3 Provide more fine grained error reason for reqwest error (#15032) 2025-12-13 13:27:06 +08:00