Commits · 69797b5ab32e4cb3742d8677a105df5b6819b220 · OpenDAS / dynamo · GitLab

19 Nov, 2025 9 commits
- feat: Only monitor NATS metrics if using NATS request plane (#4442) · 69797b5a
  Graham King authored Nov 19, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  69797b5a
- ci: TRT-LLM multimodal CI (#4118) · a8e5328e
  Indrajit Bhosale authored Nov 19, 2025
```
Signed-off-by: Indrajit Bhosale <iamindrajitb@gmail.com>
```
  a8e5328e
- chore: Upgrade NIXL to 0.7.1 and allow KVBM to pick NIXL cuda 12 or 13 (#4281) · 0b8b7ffb
  Kyle McGill authored Nov 19, 2025
```
Signed-off-by: Tushar Sharma <tusharma@nvidia.com>
Co-authored-by: Tushar Sharma <tusharma@nvidia.com>
```
  0b8b7ffb
- chore: Unpinned accelerate version (#4458) · d01b6d2d
  Kris Hung authored Nov 19, 2025
```
Signed-off-by: Kris Hung <krish@nvidia.com>
```
  d01b6d2d
- feat: nvext field to OpenAI APIs and add worker_id reporting (vLLM) (#4372) · 263f99dc
  Zhongxuan (Daniel) Wang authored Nov 18, 2025
```
Signed-off-by: Zhongxuan Wang <daniewang@nvidia.com>
```
  263f99dc
- feat: generate deterministic lora id from lora name (#4457) · 684107c4
  Biswa Panda authored Nov 18, 2025
  
  684107c4
- feat: Request Cancellation when transitioning from Prefill to Decode at KV Prefill Router (#4449) · 04486c1d
  Jacky authored Nov 18, 2025
```
Signed-off-by: Jacky <18255193+kthui@users.noreply.github.com>
```
  04486c1d
- fix: Allow users to set `--kv-transfer-config` for vLLM (#4317) · 45529d06
  Joe Chandler authored Nov 18, 2025
```
Signed-off-by: Joe Chandler <jchandler@glia-ai.com>
```
  45529d06
- docs: kvbm wheels doc fix (#4455) · 429ddf12
  Olga Andreeva authored Nov 18, 2025
```
Signed-off-by: Olga Andreeva <oandreeva@nvidia.com>
```
  429ddf12
18 Nov, 2025 22 commits
- fix: Recipe namespace fix (#4445) · c3195612
  atchernych authored Nov 18, 2025
```
Signed-off-by: Anna Tchernych <atchernych@nvidia.com>
```
  c3195612
- feat: Support toolcall parser for DeepSeek V3 and R1 (#4253) · 164b0c29
  tangcy98 authored Nov 19, 2025
```
Signed-off-by: zhangzhang <tangchenyu@xiaohongshu.com>
Co-authored-by: zhangzhang <tangchenyu@xiaohongshu.com>
Co-authored-by: Ayush Agarwal <ayushag@nvidia.com>
```
  164b0c29
- feat: Support a dynamic default max_tokens for VLLM backend (#4156) · 58405177
  bin.pan authored Nov 19, 2025
```
Signed-off-by: bin <bin.pan@daocloud.io>
```
  58405177
- fix: guard KV-indexer against self-referential blocks (#4395) · 4f2cbec0
  Vladislav Nosivskoy authored Nov 19, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
Signed-off-by: PeaBrane <yanrpei@gmail.com>
Co-authored-by: PeaBrane <yanrpei@gmail.com>
```
  4f2cbec0
- fix: respect router env vars in frontend configuration (#4431) · c65f1ac0
  Vladislav Nosivskoy authored Nov 19, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
```
  c65f1ac0
- chore: update aiperf version (#4438) · 81491b39
  Tushar Sharma authored Nov 18, 2025
```
Signed-off-by: Tushar Sharma <tusharma@nvidia.com>
```
  81491b39
- feat: support prompt_tokens_details in usage (#4239) · be9d6b2b
  Vladislav Nosivskoy authored Nov 19, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
```
  be9d6b2b
- fix: use --num-dataset-entries instead of --conversation-num (#4439) · 0f4d7634
  Anthony Casagrande authored Nov 18, 2025
  
  0f4d7634
- feat: add custom gpu type to CRD (#4408) · 9caafa55
  Thomas Montfort authored Nov 18, 2025
  
  9caafa55
- fix: Set annotated_usage.event to None in the final usage chunk response (#3729) · 33d9ae78
  Shuaiyi Zhang authored Nov 19, 2025
```
Signed-off-by: Shuaiyi Zhang <zhangsy28@lenovo.com>
Co-authored-by: Shuaiyi Zhang <zhangsy28@lenovo.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  33d9ae78
- fix: Mark enable_kv_events=True for user-specified vLLM KVEventsConfig in scripts (#4418) · 7b967c77
  Karen Chung authored Nov 18, 2025
```
Signed-off-by: Karen Chung <karenc@nvidia.com>
```
  7b967c77
- fix(storage): File store shouldn't notify on metdata changes (#4434) · 2f1778c1
  Graham King authored Nov 18, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  2f1778c1
- feat: nixl_connect: Change the way `bytes` are passed to NIXL (#4314) · 08146531
  J Wyman authored Nov 18, 2025
```
Signed-off-by: J Wyman <jwyman@nvidia.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  08146531
- fix: TENSORRTLLM_PIP_WHEEL_DIR has reasonable default set when building and... · 771d8c02
  Kyle McGill authored Nov 18, 2025
```
fix: TENSORRTLLM_PIP_WHEEL_DIR has reasonable default set when building and installing TRTLLM from source (#4414)
```
  771d8c02
- feat: mocker can use planner profile data (#4422) · fc2ad4eb
  Yan Ru Pei authored Nov 18, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
Signed-off-by: Yan Ru Pei <yanrpei@gmail.com>
Co-authored-by: Hongkuan Zhou <tedzhouhk@gmail.com>
```
  fc2ad4eb
- fix: point to new sglang container and rm references to old (#4383) · 48c340f4
  ishandhanani authored Nov 17, 2025
  
  48c340f4
- fix: add kv events config in agg_router sglang example (#4391) · cf64ca7c
  Vladislav Nosivskoy authored Nov 18, 2025
```
Signed-off-by: Vladislav Nosivskoy <vladnosiv@gmail.com>
```
  cf64ca7c
- fix: Move trtllm multimodal worker to use ModelInput.Token (#4373) · c3984bc2
  Indrajit Bhosale authored Nov 17, 2025
```
Signed-off-by: Indrajit Bhosale <iamindrajitb@gmail.com>
```
  c3984bc2
- feat: discovery traits, systems and managers (#4070) · 4d3269cd
  Ryan Olson authored Nov 17, 2025
```
Signed-off-by: Ryan Olson <rolson@nvidia.com>
```
  4d3269cd
- fix: Fix cublas and cublasLt symlinks in vLLM runtime container (#4421) · 1a27649c
  jthomson04 authored Nov 17, 2025
```
Signed-off-by: jthomson04 <jwillthomson19@gmail.com>
```
  1a27649c
- fix: Fix port collision and pass prefill param in kvbm connector (#4411) · e3cee95f
  Kris Hung authored Nov 17, 2025
```
Signed-off-by: krishung5 <krish@nvidia.com>
```
  e3cee95f
- docs: Clean up incomplete recipes and clarify Kubernetes-only focus (#4159) · 88dfd1b3
  Ben Hamm authored Nov 17, 2025
```
Signed-off-by: Ben Hamm <ben.hamm@gmail.com>
Signed-off-by: Tanmay Verma <tanmay2592@gmail.com>
Signed-off-by: atchernych <atchernych@nvidia.com>
Co-authored-by: Biswa Panda <biswa.panda@gmail.com>
Co-authored-by: tanmayv25 <tanmay2592@gmail.com>
Co-authored-by: Tanmay Verma <tanmayv@nvidia.com>
Co-authored-by: Anant Sharma <anants@nvidia.com>
Co-authored-by: atchernych <atchernych@nvidia.com>
```
  88dfd1b3
17 Nov, 2025 9 commits
- feat: kvbm kernels (#4356) · 09bb1c68
  Ryan Olson authored Nov 17, 2025
```
Signed-off-by: Ryan Olson <rolson@nvidia.com>
```
  09bb1c68
- fix: change labels to annotations in reference dgd (#4410) · adbe133c
  mohammedabdulwahhab authored Nov 17, 2025
```
Signed-off-by: mohammedabdulwahhab <furkhan324@berkeley.edu>
```
  adbe133c
- ci: Disable deploy tests (#4397) · e5f7ee9f
  Dillon Cullinan authored Nov 17, 2025
```
Signed-off-by: Dillon Cullinan <dcullinan@nvidia.com>
```
  e5f7ee9f
- fix: Install jq in trtllm runtime container (#4398) · 54addb54
  Tushar Sharma authored Nov 17, 2025
```
Signed-off-by: Tushar Sharma <tusharma@nvidia.com>
```
  54addb54
- fix: workaround vllm prometheus cleanup error on shutdown (#4323) · 6afb4d5b
  Keiven C authored Nov 17, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  6afb4d5b
- refactor: centralize environment variable constants (#4083) · 0e77d344
  Keiven C authored Nov 17, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  0e77d344
- fix: DSR1 DEP Prefill Profiling Benchmark (#4367) · cf97c0dc
  Hongkuan Zhou authored Nov 17, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  cf97c0dc
- feat: convert profiler perf data for mocker to use (#4370) · 24af5a33
  Hongkuan Zhou authored Nov 17, 2025
```
Signed-off-by: hongkuanz <hongkuanz@nvidia.com>
```
  24af5a33
- ci: Output node name in workflows (#4350) · fa28b1ac
  Dillon Cullinan authored Nov 17, 2025
```
Signed-off-by: Dillon Cullinan <dcullinan@nvidia.com>
```
  fa28b1ac