Commits · 6deeecb1d6a9f4eb1770b4272bfa85a4b6226e0a · OpenDAS / dynamo · GitLab

25 Oct, 2025 1 commit
- fix: char boundary panic jail (#3893) · 6deeecb1
  ryan-lempka authored Oct 25, 2025
  
  6deeecb1
24 Oct, 2025 7 commits
- fix: default when tools missing and minijinja length handling (#3880) · 818d72ae
  ryan-lempka authored Oct 24, 2025
```
Signed-off-by: Ryan Lempka <rlempka@nvidia.com>
Signed-off-by: ryan-lempka <rlempka@nvidia.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  818d72ae
- refactor: redesign the metrics API from Trait to composition to make the code... · cbe0b177
  Keiven C authored Oct 24, 2025
```
refactor: redesign the metrics API from Trait to composition to make the code cleaner and easier to understand (#3687)
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  cbe0b177
- chore: mocker states do not need arc mutex (#3883) · d870d4b0
  Yan Ru Pei authored Oct 24, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  d870d4b0
- chore: don't error on failure to download snapshot (#3871) · 8d0a5eba
  Yan Ru Pei authored Oct 24, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  8d0a5eba
- feat: Extend TensorRequest and TensorResponse to contains extra parameters (#3761) · 90caf3ea
  KrishnanPrash authored Oct 24, 2025
```
Signed-off-by: Krishnan Prashanth <kprashanth@nvidia.com>
Signed-off-by: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  90caf3ea
- feat: add nats sink for audit logging (#3732) · cb0a3449
  ryan-lempka authored Oct 23, 2025
```
Signed-off-by: Ryan Lempka <rlempka@nvidia.com>
```
  cb0a3449
- fix: include request_id and error details in completions stream failure message (#3860) · 2f89c176
  zhongdaor-nv authored Oct 23, 2025
```
Signed-off-by: zhongdaor <zhongdaor@nvidia.com>
```
  2f89c176
23 Oct, 2025 3 commits
- chore: Use KeyValueStoreManager instead of etcd::Client (#3822) · 7731b024
  Graham King authored Oct 23, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  7731b024
- chore: restructure mocker cli args handling, to include prefill/decode (#3847) · 41ff394f
  Yan Ru Pei authored Oct 23, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  41ff394f
- refactor: kvbm modularity DIS-657 Eliminate ETCD from the leader-worker initialization (#3202) · 94aa2a7b
  Richard Huo authored Oct 23, 2025
```
Signed-off-by: richardhuo-nv <rihuo@nvidia.com>
```
  94aa2a7b
22 Oct, 2025 4 commits
- chore: no need to publish kv events for decode vllm worker (#3819) · fb294b90
  Yan Ru Pei authored Oct 22, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  fb294b90
- fix: Load deployment card from ModelExpress for EPP (#3793) · c8adbe6f
  atchernych authored Oct 22, 2025
```
Signed-off-by: Anna Tchernych <atchernych@nvidia.com>
```
  c8adbe6f
- fix: correct prefill/decode block defaults when no overlaps (#3811) · db8e52f2
  Yan Ru Pei authored Oct 22, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  db8e52f2
- fix: enhance gRPC frontend to return output in raw content field for Triton... · d68e4b8a
  GuanLuo authored Oct 22, 2025
```
fix: enhance gRPC frontend to return output in raw content field for Triton client compatibility (#3600)
Signed-off-by: Guan Luo <gluo@nvidia.com>
Signed-off-by: Guan Luo <41310872+GuanLuo@users.noreply.github.com>
```
  d68e4b8a
21 Oct, 2025 2 commits
- feat: bake prefill router into frontend, supporting vllm for now (#3762) · e01c6e99
  Yan Ru Pei authored Oct 21, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  e01c6e99
- feat: enable KVBM GPU offload to Disk bypassing CPU (#3510) · d5840cda
  Ziqi Fan authored Oct 21, 2025
```
Signed-off-by: Ziqi Fan <ziqif@nvidia.com>
```
  d5840cda
20 Oct, 2025 3 commits
- fix: block_manager/disk - fix `fallocate` failure on some file systems (#2680) · 6ace46ce
  Dan Aloni authored Oct 21, 2025
```
Signed-off-by: Dan Aloni <dan.aloni@vastdata.com>
Co-authored-by: Ziqi Fan <ziqif@nvidia.com>
```
  6ace46ce
- fix: fixing the NUMA sensitivity problem in KVBM for TP=1 (#3700) · 67d27bcc
  Olga Andreeva authored Oct 20, 2025
```
Signed-off-by: Olga Andreeva <oandreeva@nvidia.com>
```
  67d27bcc
- chore: Replace ServiceConfigBuilder with add_stats_service (#3736) · f6ed01b1
  Graham King authored Oct 20, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  f6ed01b1
18 Oct, 2025 1 commit
- feat: add prefill workers to discovery (#3709) · 4b7a806c
  Yan Ru Pei authored Oct 17, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  4b7a806c
17 Oct, 2025 4 commits
- refactor: Make `nats_client` optional internally (#3705) · 66fd6f84
  Graham King authored Oct 17, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  66fd6f84
- chore: remove kv metrics scraping and aggregation (#3701) · 4c207e0c
  Yan Ru Pei authored Oct 17, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  4c207e0c
- feat(frontend): Get model config files (`tokenizer.json` et al.) from MX (#3659) · 9d03b8dc
  Graham King authored Oct 17, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  9d03b8dc
- feat: use non-blocking lock for radix uploading + a read lock for radix downloading (#3655) · e93170d6
  Yan Ru Pei authored Oct 16, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  e93170d6
16 Oct, 2025 4 commits
- fix: handling 0 block transfer (#3613) · 1972f71f
  Olga Andreeva authored Oct 16, 2025
```
Signed-off-by: Olga Andreeva <oandreeva@nvidia.com>
```
  1972f71f
- chore: move worker_monitor to the llm crate (#3667) · 7aa8e0e6
  Yan Ru Pei authored Oct 16, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  7aa8e0e6
- feat: Enable cancellation during or before a stream is established (#3635) · 734d2f87
  Jacky authored Oct 15, 2025
```
Signed-off-by: Jacky <18255193+kthui@users.noreply.github.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  734d2f87
- feat: dp rank routing (#3597) · f978f4d1
  Yan Ru Pei authored Oct 15, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  f978f4d1
15 Oct, 2025 2 commits
- feat: Python binding to download a model. (#3593) · ab0da582
  Graham King authored Oct 15, 2025
```
Signed-off-by: Graham King <grahamk@nvidia.com>
```
  ab0da582
- feat: router frees request from slot manager on stopped requests (#3623) · 6a1391eb
  Yan Ru Pei authored Oct 14, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  6a1391eb
14 Oct, 2025 1 commit
- fix: Add support for single element arrays for chat and completions prompts (#3482) · 3dc25129
  zhongdaor-nv authored Oct 14, 2025
```
Signed-off-by: zhongdaor <zhongdaor@nvidia.com>
```
  3dc25129
13 Oct, 2025 3 commits
- fix: Improve Error Messaging for port/address already in use error (#3587) · fa4ce46c
  KrishnanPrash authored Oct 13, 2025
```
Signed-off-by: Krishnan Prashanth <kprashanth@nvidia.com>
```
  fa4ce46c
- feat: remove stale workers on snapshot + some refactoring (#3589) · b5e762b2
  Yan Ru Pei authored Oct 13, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  b5e762b2
- feat: Metric histogram buckets configureable from env (#3562) · f0065bb4
  William Arnold authored Oct 13, 2025
```
Signed-off-by: William Arnold <7565007+Aphoh@users.noreply.github.com>
```
  f0065bb4
11 Oct, 2025 4 commits
- chore: consolidations of KvPushRouter bindings and usage examples (#3543) · c3fcfdd6
  Yan Ru Pei authored Oct 10, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
Signed-off-by: Yan Ru Pei <yanrpei@gmail.com>
Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>
```
  c3fcfdd6
- feat: implement custom backend metrics for NIM (#3266) · 65cc5337
  Keiven C authored Oct 10, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  65cc5337
- feat: rm the old KVBM metrics | update G2 to G3 metrics collection (#3561) · 4224e57d
  Ziqi Fan authored Oct 10, 2025
```
Signed-off-by: Ziqi Fan <ziqif@nvidia.com>
```
  4224e57d
- feat: add SGLang and vLLM passthrough metrics on Dynamo backend worker (#3539) · 55e458d8
  Keiven C authored Oct 10, 2025
```
Signed-off-by: Keiven Chang <keivenchang@users.noreply.github.com>
Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com>
```
  55e458d8
10 Oct, 2025 1 commit
- test: unit test for Router slot manager (#3559) · 6e8529fd
  Yan Ru Pei authored Oct 10, 2025
```
Signed-off-by: PeaBrane <yanrpei@gmail.com>
```
  6e8529fd