Proto commits in triton-inference-server/common

These 74 commits are when the Protocol Buffers files have changed:

2026-04-02

doc: Enforce `max_inflight_requests` as a shared limit across ensemble requests (#152)

The documentation is generated from this commit.

2025-11-03

feat: Add support for `max_inflight_requests` parameter to prevent unbounded memory growth in ensemble models (#141) Co-authored-by: Yingge He <157551214+yinggeh@users.noreply.github.com>

2025-03-26

Pre-Commit fix

Draft for ModelInfer

Draft for ModelInfer

2025-02-26

Create New service for callback

Create New service for callback

Create New service for callback

2024-11-06

feat: Per-model metric customization (#126)

2024-09-23

Test

2024-07-27

Fix shape and reformat free tensor handling in the input byte size check (#125) * Update model_config.proto

2024-02-23

Add cancellation into response statistics (#113)

2024-02-17

Add response statistics (#112) * Add response stats to protobuf * Remove mentioning decoupled on comments

2024-02-01

Support Double-Type Infer/Response Parameters * Support Double-Type Infer/Response Parameters

2024-01-11

Add runtime to model configuration (#103) * Add runtime to model config * Update copyright

2023-11-19

Generative ->Iterative (#107) (#108) * name change * updated language * updated with default value * updated language Co-authored-by: Neelay Shah <neelays@nvidia.com>

Generative ->Iterative (#107) * name change * updated language * updated with default value * updated language

2023-11-15

Add options for growable memory and single state buffers (#104) * Add same input/output bstate buffer option * Add an option for using GrowableMemory * Review comments * Format * Review comments * Review comment * Fix description

Add options for growable memory and single state buffers (#104) * Add same input/output bstate buffer option * Add an option for using GrowableMemory * Review comments * Format * Review comments * Review comment * Fix description

2023-11-01

Add new sequence batcher parameter for generative sequence (#102)

2023-09-06

Add GitHub action to format and lint code (#96) * Add pre-commit hook * Run commit hooks, remove ignored word list * Add GitHub action * Add Java to Clang * Fix pre-commit to include all Python files * Remove old formatter * Remove unused skipped files * Remove codeql because no more Python * Add more pre-commit filetype checkers * Trim whitespace hook * Remove unnecessary dependency * Add mixed-line-ending and case-conflicts checks * Add copyright * Update max-line-length * Remove unnecessary line * End of file * Fix comment * Add and apply isort * Remove duplicate copyrights, add hooks link * Pin workflow Ubuntu version * Flake8 Black style, move Flake8 conf to toml * Alphabetize configs by tool * Move flake8 back into pre-commit-config * Restore clang-format file * Eof newline * Fix yaml spacing * Normalize spacing * Normalize config indentation * Update line limit in clang-format to 80 chars * Update workflows to run on every PR * Run pre-commit

2023-08-15

Add preserve_ordering field to oldest strategy in sequence scheduler config (#97) (#98) Co-authored-by: Ryan McCormick <rmccormick@nvidia.com>

Add preserve_ordering field to oldest strategy in sequence scheduler config (#97)

2023-06-22

Auto-format (#95)

2023-06-08

Updating Service and Model Config Protobuf for uint64 Request Priority (#93) Co-authored-by: qmas <q.massoz@evs.com>

Revert "Updating Service and Model Config Protobuf for uint64 Request Priority" (#92) This reverts commit e3048594e2ed6d7532099c80b8fb26ec42dd7fe9.

Revert "Updating Service and Model Config Protobuf for uint64 Request Priority" This reverts commit e3048594e2ed6d7532099c80b8fb26ec42dd7fe9.

Updating Service and Model Config Protobuf for uint64 Request Priority * change priority from uint32 to uint64 in model_config * add uint64 and double types to inference parameters Co-authored-by: qmas <q.massoz@evs.com>

2023-06-07

updated with documentation on support for double and uint64

2023-05-31

adding uint64 and double param to infer parameter.

change priority from uint32 to uint64 in model_config Co-authored-by: qmas <q.massoz@evs.com>

2023-05-25

Add memory usage report in GRPC statistic service (#88) * Update GRPC service proto * Fix type * Fix type

2023-05-08

Update documentation for "platform" (#89)

2023-02-21

Add reserved namespace field in ensemble step (#81)

2023-01-23

Add protobuf for GRPC health check (#80) * Draft health service * Formatting * Clean up * Change build order * Add health proto to targets * Change ordering * Reordering build * Add comments * Copyrights, formatting * Keep implemented methods * Remove Python health executables * Rename health library * Naming

2022-10-28

Improve the documentation for input_data_file. (#76)

2022-10-11

Revert per response metrics

2022-10-10

Revert per response metrics (#74)

2022-09-20

Add response statistics to GRPC frontend (#71) * Add response statistics to GRPC frontend * Improve docs * Improve comments * Add no response count * Improve documentation clarity Co-authored-by: kthui <18255193+kthui@users.noreply.github.com>

2022-08-15

Update documentation for execution accelerators

2022-08-10

Incorporating Dynamic Logging (#70) * Migrating Changes * New line * Add comments

2022-06-15

Add 'count' field for warmup (#61) * Add 'repeat_count' field for warmup * Address comment * Change "repeat_count" to "count"

2022-05-09

Extend GRPC ModelRepositoryParameter to allow bytes (#51)

2022-04-28

Add TYPE_BF16 scaffolding (#49) * TYPE_BF16 scaffolding * Add note on BF16 datatype requiring use raw contents

2022-03-10

Add batch input item shape specification (#43) * Add batch input item shape specification * Fix copyright * Address comment

Add cache_miss to grpc stub (#42) * Add cache_miss to grpc stub * Update 2022 copyright header * Review comments

2022-02-23

Extend load API in GRPC service (#41) * Extend load API * Fix copyright

2022-02-09

Update GRPC stub to include cache stats (#37) * Add cache_hit stat to common GRPC protobuf * Update GRPC proto to match server/docs/protocol/extension_statistics.md * Add more details on cache hits per review feedback * Add more details to 'cache_hit' field and refer to it in the 'compute_*' fields

2022-02-08

Add GRPC trace service (#40) * Add GRPC trace service * Fix up * Address comment * Expose JSON null check * Address comment * Address comment

2022-01-28

map cannot be in oneof - create new message for map

fix TensorStructure def

review edits

2022-01-27

cleanup

test

Add TensorStructure field for I/O

2021-12-14

Add state initialization setting to model config protobuf (#36) * Add state initialization setting to model config protobuf * Review edit * Remove nested metadata

2021-12-10

Add optional field in ModelInput message (#35) * Add optional field in ModelInput message * Fix comment

2021-12-07

Review edit

Add state initialization setting to model config protobuf

2021-11-10

Fix GRPC protocol error. KServer protocol specifies 'bytes_contents' (#34)

2021-10-29

Add state description to model config (#28) * Add state description to the protobuf * Review edits

2021-10-08

Add response cache enable setting in model config (#30) * Add response cache enable setting in model config * Format fix * Use composite message for response cache settings

2021-10-05

Add clarification for rate limiter config priority (#29)

Remove some legacy 'custom backend' references

2021-09-24

Document memory impact of the output_copy_stream (#27)

2021-09-20

bug fix

add more configuration params for ORT

2021-08-12

Extend START, END, READY controls to allow BOOL type (#22) * Add bool type * Update identifier * Update identifier Co-authored-by: Kris Hung <krish@krish-dt.nvidia.com>

2021-05-27

Add host policy field (#17) * Add numa id field * Enforce the NUMA id to be the same as GPU id for GPU instance * Modify to "host_policy" as a more general approach * Address comment * Fix rebase artifact

2021-05-26

Add support for DLA/secondary device specification (#18) * Add support for DLA/secondary device specification * Address review comments * Improve description and other cleanup

2021-05-14

Add 'passive' field in ModelInstanceGroup (#16)

2021-05-11

Integrate minor doc changes (#15)

2021-04-16

Move protobuf to common

Integrate change from triton-inference-server/server repo > e2208d2dd5effd0 src/core/grpc_service.proto > commit 09271f9c4d4d935bd9667dd2be2208d2dd5effd0 > Author: GuanLuo <41310872+GuanLuo@users.noreply.github.com> > Date: Mon Apr 5 09:48:32 2021 -0700 > > Add end point to unload model and its dependents (#2684)