feat: http disconnects #2014
Merged
feat: http disconnects #2014
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…le; added request_id annotation and setting the request id from either request.user, x-dynamo-request-id header or server generate
WalkthroughThis update introduces a new Changes
Sequence Diagram(s)sequenceDiagram
participant Client
participant HTTPService
participant DisconnectMonitor
participant EngineContext
Client->>HTTPService: Sends request (unary or streaming)
HTTPService->>DisconnectMonitor: Create connection handles
HTTPService->>EngineContext: Start processing/generation
Note over HTTPService,DisconnectMonitor: Monitor client connection state
alt Client disconnects unexpectedly
DisconnectMonitor->>EngineContext: kill() (cancel processing)
else Client completes normally
DisconnectMonitor->>EngineContext: No action (graceful close)
end
HTTPService-->>Client: Sends response or ends stream
Poem
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
There was a problem hiding this comment.
Actionable comments posted: 0
🔭 Outside diff range comments (2)
lib/llm/src/http/service/openai.rs (2)
210-210: Remove unused request_id variablesThese request_id variables are generated but immediately overwritten or unused:
- Line 210: Overwritten by context.id()
- Line 310: Never used in embeddings handler
- Line 634: Context already has the ID
- // todo - extract distributed tracing id and context id from headers - let request_id = uuid::Uuid::new_v4().to_string(); // todo - decide on defaultAlso applies to: 310-310, 634-634
608-618: Inconsistent validation function return typesThe
validate_response_unsupported_fieldsandvalidate_response_input_is_text_onlyfunctions returnOption<impl IntoResponse>while other validation functions returnResult<(), ErrorResponse>. This inconsistency makes error handling awkward.Update these functions to return
Result<(), ErrorResponse>for consistency:- if let Some(resp) = validate_response_unsupported_fields(&request) { - return Ok(resp.into_response()); - } + validate_response_unsupported_fields(&request)?; -pub fn validate_response_unsupported_fields( - request: &NvCreateResponse, -) -> Option<impl IntoResponse> { +pub fn validate_response_unsupported_fields( + request: &NvCreateResponse, +) -> Result<(), ErrorResponse> { let inner = &request.inner; if inner.background == Some(true) { - return Some(ErrorMessage::not_implemented_error( + return Err(ErrorMessage::not_implemented_error( "`background: true` is not supported.", )); } // ... similar changes for other fields ... - None + Ok(()) }Also applies to: 708-811
🧹 Nitpick comments (3)
lib/llm/src/http/service/disconnect.rs (1)
182-184: Track the TODO for dynamo sentinel event handlingThe TODO comment indicates a potential ordering issue where a dynamo sentinel event might need to be yielded before the
[DONE]event to prevent the async-openai client from "chomping" it.Would you like me to create an issue to track this TODO item for proper sentinel event ordering?
lib/llm/tests/http-service.rs (2)
91-91: Document the reason for using deprecated max_tokensThe
#[allow(deprecated)]attribute suppresses warnings but doesn't explain whymax_tokensis still being used.Add a comment explaining why the deprecated field is still necessary:
- // ALLOW: max_tokens is deprecated in favor of completion_usage_tokens + // ALLOW: max_tokens is deprecated in favor of max_completion_tokens + // but we still use it for backward compatibility in tests #[allow(deprecated)]
1134-1135: Technical debt: Improve test fixturesThe TODO comment correctly identifies that the test setup is repetitive. Each test creates its own service, engines, and client configuration.
Would you like me to help refactor the test setup into reusable fixtures that could reduce duplication across these integration tests?
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (9)
lib/llm/src/http/service.rs(1 hunks)lib/llm/src/http/service/disconnect.rs(1 hunks)lib/llm/src/http/service/openai.rs(29 hunks)lib/llm/src/protocols.rs(3 hunks)lib/llm/src/protocols/openai/chat_completions/aggregator.rs(3 hunks)lib/llm/src/protocols/openai/completions/aggregator.rs(3 hunks)lib/llm/src/protocols/openai/embeddings/aggregator.rs(3 hunks)lib/llm/tests/http-service.rs(5 hunks)lib/runtime/src/pipeline/context.rs(1 hunks)
🧰 Additional context used
🧠 Learnings (8)
lib/llm/src/protocols.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/src/protocols/openai/embeddings/aggregator.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: t-ob
PR: ai-dynamo/dynamo#1290
File: launch/dynamo-run/src/subprocess/sglang_inc.py:80-110
Timestamp: 2025-06-03T10:17:51.711Z
Learning: The sglang `async_encode` method does not support streaming options, so collecting all embeddings before yielding is the correct approach for embedding requests.
lib/llm/src/protocols/openai/completions/aggregator.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/src/protocols/openai/chat_completions/aggregator.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/tests/http-service.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
lib/runtime/src/pipeline/context.rs (1)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
lib/llm/src/http/service/disconnect.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1093
File: lib/llm/src/block_manager/block/registry.rs:98-122
Timestamp: 2025-05-29T06:20:12.901Z
Learning: In lib/llm/src/block_manager/block/registry.rs, the background task spawned for handling unregister notifications uses detached concurrency by design. The JoinHandle is intentionally not stored as this represents a reasonable architectural tradeoff for a long-running cleanup task.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
lib/llm/src/http/service/openai.rs (8)
Learnt from: PeaBrane
PR: ai-dynamo/dynamo#1392
File: lib/llm/src/kv_router/scoring.rs:35-46
Timestamp: 2025-06-05T01:02:15.318Z
Learning: In lib/llm/src/kv_router/scoring.rs, PeaBrane prefers panic-based early failure over Result-based error handling for the worker_id() method to catch invalid data early during development.
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
Learnt from: t-ob
PR: ai-dynamo/dynamo#1290
File: launch/dynamo-run/src/subprocess/sglang_inc.py:80-110
Timestamp: 2025-06-03T10:17:51.711Z
Learning: The sglang `async_encode` method does not support streaming options, so collecting all embeddings before yielding is the correct approach for embedding requests.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:32:05.022Z
Learning: In async-nats, the "no responders" error is represented as async_nats::error::RequestErrorKind::NoResponders. Use err.downcast_ref::<async_nats::error::RequestError>() and then check req_err.kind() against RequestErrorKind::NoResponders to handle this error properly.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:32:05.022Z
Learning: In async-nats, the "no responders" error is represented as async_nats::client::RequestErrorKind::NoResponders, not async_nats::Error::NoResponders. Use err.downcast_ref::<async_nats::client::RequestError>() and then check request_err.kind() against RequestErrorKind::NoResponders.
Learnt from: ishandhanani
PR: ai-dynamo/dynamo#1626
File: lib/llm/src/preprocessor.rs:238-239
Timestamp: 2025-06-24T20:59:35.725Z
Learning: In lib/llm/src/preprocessor.rs, the `sampling_options` call in the `preprocess_request` method is placed in the common section after the match statement on `request.prompt_input_type()`, meaning it applies to both `PromptInput::Tokens` and `PromptInput::Text` request types.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
🧬 Code Graph Analysis (1)
lib/runtime/src/pipeline/context.rs (3)
lib/runtime/src/service.rs (1)
id(66-74)lib/llm/src/protocols.rs (1)
content(47-47)lib/llm/src/protocols/openai/completions.rs (1)
content(51-53)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (4)
- GitHub Check: pre-merge-rust (.)
- GitHub Check: pre-merge-rust (lib/bindings/python)
- GitHub Check: pre-merge-rust (lib/runtime/examples)
- GitHub Check: Build and Test - vllm
🔇 Additional comments (18)
lib/llm/src/http/service.rs (1)
23-23: LGTM! Clean module addition.The public
disconnectmodule addition follows the established pattern and integrates well with the existing module structure.lib/runtime/src/pipeline/context.rs (1)
78-86: LGTM! Well-designed accessor methods.The new
id()andcontent()methods provide clean, explicit access to the context's identifier and data. The implementation correctly delegates to the controller for the ID and returns a reference to the current data, following standard Rust conventions.lib/llm/src/protocols.rs (2)
22-22: LGTM! Proper import addition.Adding the
Streamtrait import alongsideStreamExtsupports the function signature generalization.
52-66: LGTM! Excellent generalization of stream handling.The function signature has been properly generalized from a concrete
DataStreamto a genericimpl Streamtrait bound, which increases flexibility while maintaining the same functionality. The implementation correctly maps over the stream items and handles both success and error cases appropriately.lib/llm/src/protocols/openai/embeddings/aggregator.rs (3)
23-23: LGTM! Consistent import update.Adding the
Streamtrait import follows the established pattern across the codebase.
61-61: LGTM! Proper stream generalization.The method signature has been correctly generalized to accept any stream implementing the required trait bounds, improving flexibility while maintaining type safety.
136-136: LGTM! Consistent API generalization.The
from_annotated_streammethod signature follows the same generalization pattern, maintaining consistency across the aggregator interface.lib/llm/src/protocols/openai/completions/aggregator.rs (3)
19-19: LGTM! Consistent import pattern.The
Streamtrait import addition aligns with the generalization changes throughout the codebase.
67-67: LGTM! Proper method signature generalization.The
applymethod signature has been correctly generalized to work with any stream implementing the required trait, maintaining consistency with other aggregators.
186-186: LGTM! Consistent API design.The
from_annotated_streammethod signature follows the established generalization pattern, completing the uniform interface across all protocol aggregators.lib/llm/src/protocols/openai/chat_completions/aggregator.rs (1)
16-16: Good API generalization with improved flexibilityThe change from concrete
DataStreamto genericimpl Streamtrait bounds is a solid improvement that:
- Reduces coupling to specific stream implementations
- Maintains backward compatibility (DataStream implements Stream)
- Uses zero-cost abstraction with static dispatch
- Aligns with Rust's principle of accepting the most general interface
Also applies to: 97-97, 262-262
lib/llm/src/http/service/disconnect.rs (1)
1-196: Well-designed disconnect monitoring implementationThe module provides a robust solution for client disconnect detection with:
- Clear separation of concerns between connection and stream lifecycle
- Proper resource cleanup via Drop trait
- Efficient concurrent monitoring using tokio::select!
- Good documentation explaining the two-phase approach
The detached task pattern (line 110) aligns with the approved architectural approach from previous learnings.
lib/llm/tests/http-service.rs (2)
57-160: Well-designed test engine for cancellation testingThe
LongRunningEngineimplementation effectively simulates long-running operations with proper cancellation detection using atomic flags and tokio::select!. The pattern of initially setting the flag to true and only clearing it on successful completion is a good defensive approach.
955-1261: Comprehensive test coverage for disconnect handlingThe three new tests provide excellent coverage for:
- Unary request cancellation on client disconnect
- Streaming request cancellation on client disconnect
- Request ID annotation propagation in SSE streams
The tests properly verify timing constraints and use appropriate assertions. Good use of unique ports to avoid conflicts.
lib/llm/src/http/service/openai.rs (4)
12-28: Excellent error handling refactoringThe restructuring of error handling with
ErrorMessagestruct andErrorResponsetype alias provides:
- Cleaner, more consistent error responses
- Proper HTTP status code mapping
- Good separation between client errors (4xx) and server errors (5xx)
- Proper downcast handling for
HttpErrorinfrom_anyhowAlso applies to: 48-131
133-156: Well-designed request ID extraction with proper fallbacksThe
get_or_create_request_idfunction implements a robust three-tier fallback:
- Primary source (e.g., user field)
- Request header
- Generate new UUID
Good UUID validation and error handling throughout.
166-198: Excellent handler refactoring with connection monitoringThe pattern of splitting handlers into public entry points and private async implementations provides:
- Clean separation of concerns
- Proper connection lifecycle management
- Consistent error handling with task spawning
- Good integration with the disconnect monitoring module
Also applies to: 358-388, 564-594
649-652: Redundant context creationThe code creates a new context wrapper around the request, but the original context already exists and should be reused.
- let request = context.map(|mut _req| { - request.inner.stream = Some(false); - request - }); + let request = context.map(|mut req| { + req.inner.stream = Some(false); + req + });⛔ Skipped due to learnings
Learnt from: ryanolson PR: ai-dynamo/dynamo#1919 File: lib/runtime/src/engine.rs:168-168 Timestamp: 2025-07-14T21:25:56.930Z Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
|
@coderabbitai full review |
✅ Actions performedFull review triggered. |
There was a problem hiding this comment.
Actionable comments posted: 0
🔭 Outside diff range comments (2)
lib/llm/src/http/service/openai.rs (2)
209-211: Remove dead code.The
request_idvariable is generated but never used, as request ID handling has been moved to the handler function.- // todo - extract distributed tracing id and context id from headers - let request_id = uuid::Uuid::new_v4().to_string();
309-311: Remove unused request_id variable.Similar to the completions function, this request_id is generated but never used.
- // todo - extract distributed tracing id and context id from headers - let request_id = uuid::Uuid::new_v4().to_string();
🧹 Nitpick comments (3)
lib/llm/tests/http-service.rs (1)
1131-1261: Thorough request ID annotation test!The test comprehensively validates request ID propagation through SSE annotations. The TODO comment about fixtures is valid - consider creating test helper functions to reduce setup boilerplate.
lib/llm/src/http/service/disconnect.rs (1)
1-30: Fix typo in documentation.Line 4: "monitoring and responding" should be "monitor and respond".
-//! The `disconnect` module provides a mechanism for our axum http services to monitoring and responding +//! The `disconnect` module provides a mechanism for our axum http services to monitor and respondlib/llm/src/http/service/openai.rs (1)
133-156: Robust request ID resolution!The function implements a good fallback chain: primary source → header → generated UUID. Consider adding debug logging when falling back to a generated UUID to help with troubleshooting.
let uuid = match request_id_opt { Some(request_id) => { - uuid::Uuid::parse_str(request_id).unwrap_or_else(|_| uuid::Uuid::new_v4()) + uuid::Uuid::parse_str(request_id).unwrap_or_else(|_| { + tracing::debug!("Invalid UUID in header, generating new one"); + uuid::Uuid::new_v4() + }) } - None => uuid::Uuid::new_v4(), + None => { + tracing::debug!("No request ID provided, generating new one"); + uuid::Uuid::new_v4() + }, };
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (9)
lib/llm/src/http/service.rs(1 hunks)lib/llm/src/http/service/disconnect.rs(1 hunks)lib/llm/src/http/service/openai.rs(29 hunks)lib/llm/src/protocols.rs(3 hunks)lib/llm/src/protocols/openai/chat_completions/aggregator.rs(3 hunks)lib/llm/src/protocols/openai/completions/aggregator.rs(3 hunks)lib/llm/src/protocols/openai/embeddings/aggregator.rs(3 hunks)lib/llm/tests/http-service.rs(5 hunks)lib/runtime/src/pipeline/context.rs(1 hunks)
🧰 Additional context used
🧠 Learnings (8)
lib/runtime/src/pipeline/context.rs (1)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
lib/llm/src/protocols/openai/embeddings/aggregator.rs (1)
Learnt from: t-ob
PR: ai-dynamo/dynamo#1290
File: launch/dynamo-run/src/subprocess/sglang_inc.py:80-110
Timestamp: 2025-06-03T10:17:51.711Z
Learning: The sglang `async_encode` method does not support streaming options, so collecting all embeddings before yielding is the correct approach for embedding requests.
lib/llm/src/protocols/openai/completions/aggregator.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/src/protocols.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/src/protocols/openai/chat_completions/aggregator.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
lib/llm/src/http/service/disconnect.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1093
File: lib/llm/src/block_manager/block/registry.rs:98-122
Timestamp: 2025-05-29T06:20:12.901Z
Learning: In lib/llm/src/block_manager/block/registry.rs, the background task spawned for handling unregister notifications uses detached concurrency by design. The JoinHandle is intentionally not stored as this represents a reasonable architectural tradeoff for a long-running cleanup task.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
lib/llm/tests/http-service.rs (2)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
lib/llm/src/http/service/openai.rs (8)
Learnt from: ryanolson
PR: ai-dynamo/dynamo#1919
File: lib/runtime/src/engine.rs:168-168
Timestamp: 2025-07-14T21:25:56.930Z
Learning: The AsyncEngineContextProvider trait in lib/runtime/src/engine.rs was intentionally changed from `Send + Sync + Debug` to `Send + Debug` because the Sync bound was overly constraining. The trait should only require Send + Debug as designed.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:07:24.843Z
Learning: The codebase uses async-nats version 0.40, not the older nats crate. Error handling should use async_nats::error::Error variants, not nats::Error variants.
Learnt from: PeaBrane
PR: ai-dynamo/dynamo#1285
File: lib/llm/src/kv_router/scoring.rs:58-63
Timestamp: 2025-05-30T06:38:09.630Z
Learning: In lib/llm/src/kv_router/scoring.rs, the user prefers to keep the panic behavior when calculating load_avg and variance with empty endpoints rather than adding guards for division by zero. They want the code to fail fast on this error condition.
Learnt from: t-ob
PR: ai-dynamo/dynamo#1290
File: launch/dynamo-run/src/subprocess/sglang_inc.py:80-110
Timestamp: 2025-06-03T10:17:51.711Z
Learning: The sglang `async_encode` method does not support streaming options, so collecting all embeddings before yielding is the correct approach for embedding requests.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:32:05.022Z
Learning: In async-nats, the "no responders" error is represented as async_nats::error::RequestErrorKind::NoResponders. Use err.downcast_ref::<async_nats::error::RequestError>() and then check req_err.kind() against RequestErrorKind::NoResponders to handle this error properly.
Learnt from: kthui
PR: ai-dynamo/dynamo#1424
File: lib/runtime/src/pipeline/network/egress/push_router.rs:204-209
Timestamp: 2025-06-13T22:32:05.022Z
Learning: In async-nats, the "no responders" error is represented as async_nats::client::RequestErrorKind::NoResponders, not async_nats::Error::NoResponders. Use err.downcast_ref::<async_nats::client::RequestError>() and then check request_err.kind() against RequestErrorKind::NoResponders.
Learnt from: ishandhanani
PR: ai-dynamo/dynamo#1626
File: lib/llm/src/preprocessor.rs:238-239
Timestamp: 2025-06-24T20:59:35.725Z
Learning: In lib/llm/src/preprocessor.rs, the `sampling_options` call in the `preprocess_request` method is placed in the common section after the match statement on `request.prompt_input_type()`, meaning it applies to both `PromptInput::Tokens` and `PromptInput::Text` request types.
Learnt from: oandreeva-nv
PR: ai-dynamo/dynamo#1195
File: lib/llm/tests/block_manager.rs:150-152
Timestamp: 2025-06-02T19:37:27.666Z
Learning: In Rust/Tokio applications, when background tasks use channels for communication, dropping the sender automatically signals task termination when the receiver gets `None`. The `start_batching_publisher` function in `lib/llm/tests/block_manager.rs` demonstrates this pattern: when the `KVBMDynamoRuntimeComponent` is dropped, its `batch_tx` sender is dropped, causing `rx.recv()` to return `None`, which triggers cleanup and task termination.
🧬 Code Graph Analysis (2)
lib/runtime/src/pipeline/context.rs (3)
lib/runtime/src/service.rs (1)
id(66-74)lib/llm/src/protocols.rs (1)
content(47-47)lib/llm/src/protocols/openai/completions.rs (1)
content(51-53)
lib/llm/src/http/service/disconnect.rs (2)
lib/runtime/src/pipeline/context.rs (8)
id(79-81)id(272-274)id(342-344)id(357-359)context(226-228)context(306-308)stopped(296-298)stopped(369-375)lib/llm/src/http/service/metrics.rs (1)
mark_ok(326-328)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)
- GitHub Check: pre-merge-rust (.)
- GitHub Check: pre-merge-rust (lib/bindings/python)
- GitHub Check: Build and Test - vllm
🔇 Additional comments (27)
lib/llm/src/http/service.rs (1)
23-23: LGTM: Clean module integrationThe public module declaration properly exposes the disconnect handling functionality to the HTTP service layer, enabling connection monitoring and cancellation features.
lib/runtime/src/pipeline/context.rs (1)
78-86: LGTM: Clean accessor methods enhance context usabilityThe new
id()andcontent()accessor methods provide explicit, well-named access to the context's identifier and wrapped data. These methods support the disconnect handling functionality by facilitating request ID extraction and data access in HTTP handlers.lib/llm/src/protocols.rs (2)
22-22: LGTM: Stream trait import for generalizationAdding the
Streamtrait import supports the generalization of stream handling functions.
52-66: LGTM: Excellent stream type generalizationThe function signature generalization from concrete
DataStreamtoimpl Streamimproves flexibility while maintaining the same functionality. The direct return of the mapped stream without boxing is more efficient and cleaner.lib/llm/src/protocols/openai/completions/aggregator.rs (3)
19-19: LGTM: Stream trait import for consistencyAdding the
Streamtrait import maintains consistency with the generalization pattern across protocol aggregators.
67-67: LGTM: Stream type generalization enhances flexibilityGeneralizing the
applymethod parameter from concreteDataStreamtoimpl Streamimproves flexibility while maintaining the same aggregation logic.
186-186: LGTM: Consistent stream generalizationThe
from_annotated_streammethod generalization aligns with the pattern established in other protocol aggregators, maintaining consistency across the codebase.lib/llm/src/protocols/openai/embeddings/aggregator.rs (3)
23-23: LGTM: Stream trait import completes generalization patternAdding the
Streamtrait import maintains consistency with the generalization pattern across all protocol aggregators.
61-61: LGTM: Stream type generalization for embeddingsGeneralizing the
applymethod parameter from concreteDataStreamtoimpl Streammaintains consistency with other protocol aggregators while improving flexibility.
136-136: LGTM: Consistent stream interface generalizationThe
from_annotated_streammethod generalization completes the consistent pattern across all protocol aggregators, enhancing flexibility for stream handling.lib/llm/src/protocols/openai/chat_completions/aggregator.rs (3)
16-16: LGTM!The import addition is correct and necessary for the generic
Streamtrait bounds used in the updated function signatures.
96-98: Good API improvement!Generalizing the stream parameter to accept any
impl Streammakes the function more flexible and follows best practices by accepting trait bounds rather than concrete types.
261-263: Consistent API improvement!The generalization to
impl Streammaintains consistency with theapplymethod and improves API flexibility.lib/llm/tests/http-service.rs (5)
31-32: LGTM!All new imports are properly used in the test implementations.
Also applies to: 50-52
57-73: Well-designed test utility!The
LongRunningEngineprovides a good way to test cancellation behavior with proper atomic operations for thread-safe cancellation tracking.
114-160: Clever cancellation detection design!The implementation correctly handles cancellation by defaulting to
cancelled=trueand only setting it tofalseon successful completion. Thetokio::select!properly monitors both the delay and context cancellation.
955-1037: Comprehensive disconnect test for unary requests!The test properly simulates client disconnect using timeout and verifies that cancellation propagates to the engine within the expected timeframe.
1039-1129: Well-designed streaming disconnect test!The test correctly simulates streaming client disconnect by dropping the stream after reading one chunk and verifies timely cancellation propagation.
lib/llm/src/http/service/disconnect.rs (4)
38-92: Well-designed connection handle abstraction!The
ConnectionHandleprovides a clean API for monitoring connection state with proper drop semantics. The multiple factory methods offer good ergonomics for different use cases.
94-121: Good connection monitoring setup!The function correctly creates two handles with appropriate initial states and spawns the monitoring task. The detached concurrency pattern is appropriate here as the task naturally terminates when both handles are dropped.
123-151: Robust connection monitoring logic!The function correctly handles all connection states and appropriately triggers cancellation via
engine_context.kill()for unexpected closures.
153-196: Excellent stream monitoring implementation!The function properly handles all cases: normal completion, errors, and context cancellation. The use of
tokio::select!ensures responsive cancellation detection.Note the TODO about yielding a dynamo sentinel event - this might be needed for proper client compatibility.
lib/llm/src/http/service/openai.rs (5)
12-47: Clean imports and well-defined constants!The new imports support the refactored functionality, and the constants for request ID handling are properly defined.
48-131: Excellent error handling refactoring!The separation of
ErrorMessagestruct fromErrorResponsetype alias improves code clarity and makes error handling more consistent throughout the module.
166-198: Excellent handler refactoring with disconnect monitoring!The split between the public handler and private async function improves code organization. The connection monitoring is properly integrated with correct handle lifecycle management.
238-269: Well-implemented annotation handling!The code correctly filters for request ID annotations and prepends them to the response stream using a clean functional approach.
358-388: Consistent handler pattern across endpoints!The refactoring pattern is consistently applied to chat_completions and responses handlers, maintaining good code organization and connection monitoring throughout.
Also applies to: 564-594
grahamking
reviewed
Jul 18, 2025
grahamking
approved these changes
Jul 18, 2025
ZichengMa
added a commit
that referenced
this pull request
Jul 21, 2025
commit cb6de94 Author: ptarasiewiczNV <104908264+ptarasiewiczNV@users.noreply.github.com> Date: Sun Jul 20 22:34:50 2025 +0200 chore: Install vLLM and WideEP kernels in vLLM runtime container (#2010) Signed-off-by: Alec <35311602+alec-flowers@users.noreply.github.com> Co-authored-by: Alec <35311602+alec-flowers@users.noreply.github.com> Co-authored-by: alec-flowers <aflowers@nvidia.com> commit fe63c17 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Fri Jul 18 17:45:08 2025 -0700 fix: Revert "feat: add vLLM v1 multi-modal example. Add llama4 Maverick ex… (#2017) commit bf1998f Author: jthomson04 <jwillthomson19@gmail.com> Date: Fri Jul 18 17:23:50 2025 -0700 fix: Don't detokenize twice in TRT-LLM examples (#1955) commit 343a481 Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Fri Jul 18 16:22:43 2025 -0600 feat: http disconnects (#2014) commit e330d96 Author: Yan Ru Pei <yanrpei@gmail.com> Date: Fri Jul 18 13:40:54 2025 -0700 feat: enable / disable chunked prefill for mockers (#2015) Signed-off-by: Yan Ru Pei <yanrpei@gmail.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> commit 353146e Author: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Date: Fri Jul 18 13:33:36 2025 -0700 feat: add vLLM v1 multi-modal example. Add llama4 Maverick example (#1990) Signed-off-by: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Co-authored-by: krishung5 <krish@nvidia.com> commit 1f07dab Author: Jacky <18255193+kthui@users.noreply.github.com> Date: Fri Jul 18 13:04:20 2025 -0700 feat: Add migration to LLM requests (#1930) commit 5f17918 Author: Tanmay Verma <tanmayv@nvidia.com> Date: Fri Jul 18 12:59:34 2025 -0700 refactor: Migrate to new UX2 for python launch (#2003) commit fc12436 Author: Graham King <grahamk@nvidia.com> Date: Fri Jul 18 14:52:57 2025 -0400 feat(frontend): router-mode settings (#2001) commit dc75cf1 Author: ptarasiewiczNV <104908264+ptarasiewiczNV@users.noreply.github.com> Date: Fri Jul 18 18:47:28 2025 +0200 chore: Move NIXL repo clone to Dockerfiles (#2009) commit f6f392c Author: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> Date: Thu Jul 17 18:44:17 2025 -0700 Remove link to the fix for disagg + eagle3 for TRT-LLM example (#2006) Signed-off-by: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> commit cc90ca6 Author: atchernych <atchernych@nvidia.com> Date: Thu Jul 17 18:34:40 2025 -0700 feat: Create a convenience script to uninstall Dynamo Deploy CRDs (#1933) commit 267b422 Author: Greg Clark <grclark@nvidia.com> Date: Thu Jul 17 20:44:21 2025 -0400 chore: loosed python requirement versions (#1998) Signed-off-by: Greg Clark <grclark@nvidia.com> commit b8474e5 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Thu Jul 17 16:35:05 2025 -0700 chore: update cmake and gap installation and sgl in wideep container (#1991) commit 157a3b0 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 15:38:12 2025 -0700 fix: incorrect helm upgrade command (#2000) commit 0dfca2c Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 15:33:33 2025 -0700 ci: Update trtllm gitlab triggers for new components directory and test script (#1992) commit f3fb09e Author: Kris Hung <krish@nvidia.com> Date: Thu Jul 17 14:59:59 2025 -0700 fix: Fix syntax for tokio-console (#1997) commit dacffb8 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 14:57:10 2025 -0700 fix: use non-dev golang image for operator (#1993) commit 2b29a0a Author: zaristei <zaristei@berkeley.edu> Date: Thu Jul 17 13:10:42 2025 -0700 fix: Working Arm Build Dockerfile for Vllm_v1 (#1844) commit 2430d89 Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 12:57:46 2025 -0700 test: Add trtllm kv router tests (#1988) commit 1eadc01 Author: Graham King <grahamk@nvidia.com> Date: Thu Jul 17 15:07:41 2025 -0400 feat(runtime): Support tokio-console (#1986) commit b62e633 Author: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Date: Thu Jul 17 11:16:28 2025 -0700 feat: support separate chat_template.jinja file (#1853) commit 8ae3719 Author: Hongkuan Zhou <tedzhouhk@gmail.com> Date: Thu Jul 17 11:12:35 2025 -0700 chore: add some details to dynamo deploy quickstart and fix deploy.sh (#1978) Signed-off-by: Hongkuan Zhou <tedzhouhk@gmail.com> Co-authored-by: julienmancuso <161955438+julienmancuso@users.noreply.github.com> commit 08891ff Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 10:57:42 2025 -0700 fix: Update trtllm tests to use new scripts instead of dynamo serve (#1979) commit 49b7a0d Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Thu Jul 17 08:35:04 2025 -0600 feat: record + analyze logprobs (#1957) commit 6d2be14 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 00:17:58 2025 -0700 refactor: replace vllm with vllm_v1 container (#1953) Co-authored-by: alec-flowers <aflowers@nvidia.com> commit 4d2a31a Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Wed Jul 16 18:04:09 2025 -0700 chore: add port reservation to utils (#1980) commit 1e3e4a0 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Wed Jul 16 15:54:04 2025 -0700 fix: port race condition through deterministic ports (#1937) commit 4ad281f Author: Tanmay Verma <tanmayv@nvidia.com> Date: Wed Jul 16 14:33:51 2025 -0700 refactor: Move TRTLLM example to the component/backends (#1976) commit 57d24a1 Author: Misha Chornyi <99709299+mc-nv@users.noreply.github.com> Date: Wed Jul 16 14:10:24 2025 -0700 build: Removing shell configuration violations. It's bad practice to hardcod… (#1973) commit 182d3b5 Author: Graham King <grahamk@nvidia.com> Date: Wed Jul 16 16:12:40 2025 -0400 chore(bindings): Remove mistralrs / llama.cpp (#1970) commit def6eaa Author: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Date: Wed Jul 16 15:50:23 2025 -0400 feat: attributions for debian deps of sglang, trtllm, vllm runtime containers (#1971) commit f31732a Author: Yan Ru Pei <yanrpei@gmail.com> Date: Wed Jul 16 11:22:15 2025 -0700 feat: integrate mocker with dynamo-run and python cli (#1927) commit aba6099 Author: Graham King <grahamk@nvidia.com> Date: Wed Jul 16 12:26:32 2025 -0400 perf(router): Remove lock from router hot path (#1963) commit b212103 Author: Hongkuan Zhou <tedzhouhk@gmail.com> Date: Wed Jul 16 08:55:33 2025 -0700 docs: add notes in docs to deprecate local connector (#1959) commit 7b325ee Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 18:52:00 2025 -0700 fix: vllm router examples (#1942) commit a50be1a Author: hhzhang16 <54051230+hhzhang16@users.noreply.github.com> Date: Tue Jul 15 17:58:01 2025 -0700 feat: update CODEOWNERS (#1926) commit e260fdf Author: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Date: Tue Jul 15 18:49:21 2025 -0400 feat: add bitnami helm chart attribution (#1943) Signed-off-by: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> commit 1c03404 Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 14:26:24 2025 -0700 fix: update inference gateway deployment instructions (#1940) commit 5ca570f Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:54:03 2025 -0400 chore: Rename dynamo.ingress to dynamo.frontend (#1944) commit 7b9182f Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:33:07 2025 -0400 chore: Move examples/cli to lib/bindings/examples/cli (#1952) commit 40d40dd Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:02:19 2025 -0400 chore(multi-modal): Rename frontend.py to web.py (#1951) commit a9e0891 Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Tue Jul 15 12:30:30 2025 -0600 feat: adding http clients and recorded response stream (#1919) commit 4128d58 Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 10:30:47 2025 -0700 feat: allow helm upgrade using deploy script (#1936) commit 4da078b Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 12:57:38 2025 -0400 fix: Remove OpenSSL dependency, use Rust TLS (#1945) commit fc004d4 Author: jthomson04 <jwillthomson19@gmail.com> Date: Tue Jul 15 08:45:42 2025 -0700 fix: Fix TRT-LLM container build when using a custom pip wheel (#1825) commit 3c6fc6f Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 22:35:20 2025 -0700 chore: fix typo (#1938) commit de7fe38 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Mon Jul 14 21:47:12 2025 -0700 feat: add vllm e2e integration tests (#1935) commit 860f3f7 Author: Keiven C <213854356+keivenchang@users.noreply.github.com> Date: Mon Jul 14 21:44:19 2025 -0700 chore: metrics endpoint variables renamed from HTTP_SERVER->SYSTEM (#1934) Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com> commit fc402a3 Author: Biswa Panda <biswa.panda@gmail.com> Date: Mon Jul 14 21:21:20 2025 -0700 feat: configurable namespace for vllm v1 example (#1909) commit df40d2c Author: ZichengMa <zichengma1225@gmail.com> Date: Mon Jul 14 21:11:29 2025 -0700 docs: fix typo and add mount-workspace to vllm doc (#1931) Signed-off-by: ZichengMa <zichengma1225@gmail.com> Co-authored-by: Alec <35311602+alec-flowers@users.noreply.github.com> commit 901715b Author: Tanmay Verma <tanmayv@nvidia.com> Date: Mon Jul 14 20:14:51 2025 -0700 refactor: Refactor the TRTLLM examples remove dynamo SDK (#1884) commit 5bf23d5 Author: hhzhang16 <54051230+hhzhang16@users.noreply.github.com> Date: Mon Jul 14 18:29:19 2025 -0700 feat: update DynamoGraphDeployments for vllm_v1 (#1890) Co-authored-by: mohammedabdulwahhab <furkhan324@berkeley.edu> commit 9e76590 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 17:29:56 2025 -0700 docs: organize sglang readme (#1910) commit ef59ac8 Author: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com> Date: Mon Jul 14 16:16:44 2025 -0700 docs: TRTLLM Example of Llama4+Eagle3 (Speculative Decoding) (#1828) Signed-off-by: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com> Co-authored-by: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> commit 053041e Author: Jorge António <matroid@outlook.com> Date: Tue Jul 15 00:06:38 2025 +0100 fix: resolve incorrect finish reason propagation (#1857) commit 3733f58 Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 19:04:22 2025 -0400 feat(backends): Python llama.cpp engine (#1925) commit 6a1350c Author: Tushar Sharma <tusharma@nvidia.com> Date: Mon Jul 14 14:56:36 2025 -0700 build: minor improvements to sglang dockerfile (#1917) commit e2a619b Author: Neelay Shah <neelays@nvidia.com> Date: Mon Jul 14 14:52:53 2025 -0700 fix: remove environment variable passing (#1911) Signed-off-by: Neelay Shah <neelays@nvidia.com> Co-authored-by: Neelay Shah <neelays@a4u8g-0057.ipp2u2.colossus.nvidia.com> commit 3d17a49 Author: Schwinn Saereesitthipitak <17022745+galletas1712@users.noreply.github.com> Date: Mon Jul 14 14:41:56 2025 -0700 refactor: remove dynamo build (#1778) Signed-off-by: Schwinn Saereesitthipitak <17022745+galletas1712@users.noreply.github.com> commit 3e0cb07 Author: Anant Sharma <anants@nvidia.com> Date: Mon Jul 14 15:43:48 2025 -0400 fix: copy attributions and license to trtllm runtime container (#1916) commit fc36bf5 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 12:31:49 2025 -0700 feat: receive kvmetrics from sglang scheduler (#1789) Co-authored-by: zixuanzhang226 <zixuanzhang@bytedance.com> commit df91fce Author: Yan Ru Pei <yanrpei@gmail.com> Date: Mon Jul 14 12:24:04 2025 -0700 feat: prefill aware routing (#1895) commit ad8ad66 Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 15:20:35 2025 -0400 feat: Shrink the ai-dynamo wheel by 35 MiB (#1918) Remove http and llmctl binaries. They have been unused for a while. commit 480b41d Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 15:06:45 2025 -0400 feat: Python frontend / ingress node (#1912)
ZichengMa
added a commit
that referenced
this pull request
Jul 21, 2025
commit d4b5414 Author: atchernych <atchernych@nvidia.com> Date: Mon Jul 21 13:10:24 2025 -0700 fix: mypy error (#2029) commit 79337c7 Author: Ryan McCormick <rmccormick@nvidia.com> Date: Mon Jul 21 12:12:16 2025 -0700 build: support custom TRTLLM build for commits not on main branch (#2021) commit 95dd942 Author: atchernych <atchernych@nvidia.com> Date: Mon Jul 21 12:09:33 2025 -0700 docs: Post-Merge cleanup of the deploy documentation (#1922) commit cb6de94 Author: ptarasiewiczNV <104908264+ptarasiewiczNV@users.noreply.github.com> Date: Sun Jul 20 22:34:50 2025 +0200 chore: Install vLLM and WideEP kernels in vLLM runtime container (#2010) Signed-off-by: Alec <35311602+alec-flowers@users.noreply.github.com> Co-authored-by: Alec <35311602+alec-flowers@users.noreply.github.com> Co-authored-by: alec-flowers <aflowers@nvidia.com> commit fe63c17 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Fri Jul 18 17:45:08 2025 -0700 fix: Revert "feat: add vLLM v1 multi-modal example. Add llama4 Maverick ex… (#2017) commit bf1998f Author: jthomson04 <jwillthomson19@gmail.com> Date: Fri Jul 18 17:23:50 2025 -0700 fix: Don't detokenize twice in TRT-LLM examples (#1955) commit 343a481 Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Fri Jul 18 16:22:43 2025 -0600 feat: http disconnects (#2014) commit e330d96 Author: Yan Ru Pei <yanrpei@gmail.com> Date: Fri Jul 18 13:40:54 2025 -0700 feat: enable / disable chunked prefill for mockers (#2015) Signed-off-by: Yan Ru Pei <yanrpei@gmail.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> commit 353146e Author: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Date: Fri Jul 18 13:33:36 2025 -0700 feat: add vLLM v1 multi-modal example. Add llama4 Maverick example (#1990) Signed-off-by: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Co-authored-by: krishung5 <krish@nvidia.com> commit 1f07dab Author: Jacky <18255193+kthui@users.noreply.github.com> Date: Fri Jul 18 13:04:20 2025 -0700 feat: Add migration to LLM requests (#1930) commit 5f17918 Author: Tanmay Verma <tanmayv@nvidia.com> Date: Fri Jul 18 12:59:34 2025 -0700 refactor: Migrate to new UX2 for python launch (#2003) commit fc12436 Author: Graham King <grahamk@nvidia.com> Date: Fri Jul 18 14:52:57 2025 -0400 feat(frontend): router-mode settings (#2001) commit dc75cf1 Author: ptarasiewiczNV <104908264+ptarasiewiczNV@users.noreply.github.com> Date: Fri Jul 18 18:47:28 2025 +0200 chore: Move NIXL repo clone to Dockerfiles (#2009) commit f6f392c Author: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> Date: Thu Jul 17 18:44:17 2025 -0700 Remove link to the fix for disagg + eagle3 for TRT-LLM example (#2006) Signed-off-by: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> commit cc90ca6 Author: atchernych <atchernych@nvidia.com> Date: Thu Jul 17 18:34:40 2025 -0700 feat: Create a convenience script to uninstall Dynamo Deploy CRDs (#1933) commit 267b422 Author: Greg Clark <grclark@nvidia.com> Date: Thu Jul 17 20:44:21 2025 -0400 chore: loosed python requirement versions (#1998) Signed-off-by: Greg Clark <grclark@nvidia.com> commit b8474e5 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Thu Jul 17 16:35:05 2025 -0700 chore: update cmake and gap installation and sgl in wideep container (#1991) commit 157a3b0 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 15:38:12 2025 -0700 fix: incorrect helm upgrade command (#2000) commit 0dfca2c Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 15:33:33 2025 -0700 ci: Update trtllm gitlab triggers for new components directory and test script (#1992) commit f3fb09e Author: Kris Hung <krish@nvidia.com> Date: Thu Jul 17 14:59:59 2025 -0700 fix: Fix syntax for tokio-console (#1997) commit dacffb8 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 14:57:10 2025 -0700 fix: use non-dev golang image for operator (#1993) commit 2b29a0a Author: zaristei <zaristei@berkeley.edu> Date: Thu Jul 17 13:10:42 2025 -0700 fix: Working Arm Build Dockerfile for Vllm_v1 (#1844) commit 2430d89 Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 12:57:46 2025 -0700 test: Add trtllm kv router tests (#1988) commit 1eadc01 Author: Graham King <grahamk@nvidia.com> Date: Thu Jul 17 15:07:41 2025 -0400 feat(runtime): Support tokio-console (#1986) commit b62e633 Author: GuanLuo <41310872+GuanLuo@users.noreply.github.com> Date: Thu Jul 17 11:16:28 2025 -0700 feat: support separate chat_template.jinja file (#1853) commit 8ae3719 Author: Hongkuan Zhou <tedzhouhk@gmail.com> Date: Thu Jul 17 11:12:35 2025 -0700 chore: add some details to dynamo deploy quickstart and fix deploy.sh (#1978) Signed-off-by: Hongkuan Zhou <tedzhouhk@gmail.com> Co-authored-by: julienmancuso <161955438+julienmancuso@users.noreply.github.com> commit 08891ff Author: Ryan McCormick <rmccormick@nvidia.com> Date: Thu Jul 17 10:57:42 2025 -0700 fix: Update trtllm tests to use new scripts instead of dynamo serve (#1979) commit 49b7a0d Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Thu Jul 17 08:35:04 2025 -0600 feat: record + analyze logprobs (#1957) commit 6d2be14 Author: Biswa Panda <biswa.panda@gmail.com> Date: Thu Jul 17 00:17:58 2025 -0700 refactor: replace vllm with vllm_v1 container (#1953) Co-authored-by: alec-flowers <aflowers@nvidia.com> commit 4d2a31a Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Wed Jul 16 18:04:09 2025 -0700 chore: add port reservation to utils (#1980) commit 1e3e4a0 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Wed Jul 16 15:54:04 2025 -0700 fix: port race condition through deterministic ports (#1937) commit 4ad281f Author: Tanmay Verma <tanmayv@nvidia.com> Date: Wed Jul 16 14:33:51 2025 -0700 refactor: Move TRTLLM example to the component/backends (#1976) commit 57d24a1 Author: Misha Chornyi <99709299+mc-nv@users.noreply.github.com> Date: Wed Jul 16 14:10:24 2025 -0700 build: Removing shell configuration violations. It's bad practice to hardcod… (#1973) commit 182d3b5 Author: Graham King <grahamk@nvidia.com> Date: Wed Jul 16 16:12:40 2025 -0400 chore(bindings): Remove mistralrs / llama.cpp (#1970) commit def6eaa Author: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Date: Wed Jul 16 15:50:23 2025 -0400 feat: attributions for debian deps of sglang, trtllm, vllm runtime containers (#1971) commit f31732a Author: Yan Ru Pei <yanrpei@gmail.com> Date: Wed Jul 16 11:22:15 2025 -0700 feat: integrate mocker with dynamo-run and python cli (#1927) commit aba6099 Author: Graham King <grahamk@nvidia.com> Date: Wed Jul 16 12:26:32 2025 -0400 perf(router): Remove lock from router hot path (#1963) commit b212103 Author: Hongkuan Zhou <tedzhouhk@gmail.com> Date: Wed Jul 16 08:55:33 2025 -0700 docs: add notes in docs to deprecate local connector (#1959) commit 7b325ee Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 18:52:00 2025 -0700 fix: vllm router examples (#1942) commit a50be1a Author: hhzhang16 <54051230+hhzhang16@users.noreply.github.com> Date: Tue Jul 15 17:58:01 2025 -0700 feat: update CODEOWNERS (#1926) commit e260fdf Author: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Date: Tue Jul 15 18:49:21 2025 -0400 feat: add bitnami helm chart attribution (#1943) Signed-off-by: Harrison Saturley-Hall <454891+saturley-hall@users.noreply.github.com> Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com> commit 1c03404 Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 14:26:24 2025 -0700 fix: update inference gateway deployment instructions (#1940) commit 5ca570f Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:54:03 2025 -0400 chore: Rename dynamo.ingress to dynamo.frontend (#1944) commit 7b9182f Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:33:07 2025 -0400 chore: Move examples/cli to lib/bindings/examples/cli (#1952) commit 40d40dd Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 16:02:19 2025 -0400 chore(multi-modal): Rename frontend.py to web.py (#1951) commit a9e0891 Author: Ryan Olson <ryanolson@users.noreply.github.com> Date: Tue Jul 15 12:30:30 2025 -0600 feat: adding http clients and recorded response stream (#1919) commit 4128d58 Author: Biswa Panda <biswa.panda@gmail.com> Date: Tue Jul 15 10:30:47 2025 -0700 feat: allow helm upgrade using deploy script (#1936) commit 4da078b Author: Graham King <grahamk@nvidia.com> Date: Tue Jul 15 12:57:38 2025 -0400 fix: Remove OpenSSL dependency, use Rust TLS (#1945) commit fc004d4 Author: jthomson04 <jwillthomson19@gmail.com> Date: Tue Jul 15 08:45:42 2025 -0700 fix: Fix TRT-LLM container build when using a custom pip wheel (#1825) commit 3c6fc6f Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 22:35:20 2025 -0700 chore: fix typo (#1938) commit de7fe38 Author: Alec <35311602+alec-flowers@users.noreply.github.com> Date: Mon Jul 14 21:47:12 2025 -0700 feat: add vllm e2e integration tests (#1935) commit 860f3f7 Author: Keiven C <213854356+keivenchang@users.noreply.github.com> Date: Mon Jul 14 21:44:19 2025 -0700 chore: metrics endpoint variables renamed from HTTP_SERVER->SYSTEM (#1934) Co-authored-by: Keiven Chang <keivenchang@users.noreply.github.com> commit fc402a3 Author: Biswa Panda <biswa.panda@gmail.com> Date: Mon Jul 14 21:21:20 2025 -0700 feat: configurable namespace for vllm v1 example (#1909) commit df40d2c Author: ZichengMa <zichengma1225@gmail.com> Date: Mon Jul 14 21:11:29 2025 -0700 docs: fix typo and add mount-workspace to vllm doc (#1931) Signed-off-by: ZichengMa <zichengma1225@gmail.com> Co-authored-by: Alec <35311602+alec-flowers@users.noreply.github.com> commit 901715b Author: Tanmay Verma <tanmayv@nvidia.com> Date: Mon Jul 14 20:14:51 2025 -0700 refactor: Refactor the TRTLLM examples remove dynamo SDK (#1884) commit 5bf23d5 Author: hhzhang16 <54051230+hhzhang16@users.noreply.github.com> Date: Mon Jul 14 18:29:19 2025 -0700 feat: update DynamoGraphDeployments for vllm_v1 (#1890) Co-authored-by: mohammedabdulwahhab <furkhan324@berkeley.edu> commit 9e76590 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 17:29:56 2025 -0700 docs: organize sglang readme (#1910) commit ef59ac8 Author: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com> Date: Mon Jul 14 16:16:44 2025 -0700 docs: TRTLLM Example of Llama4+Eagle3 (Speculative Decoding) (#1828) Signed-off-by: KrishnanPrash <140860868+KrishnanPrash@users.noreply.github.com> Co-authored-by: Iman Tabrizian <10105175+Tabrizian@users.noreply.github.com> commit 053041e Author: Jorge António <matroid@outlook.com> Date: Tue Jul 15 00:06:38 2025 +0100 fix: resolve incorrect finish reason propagation (#1857) commit 3733f58 Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 19:04:22 2025 -0400 feat(backends): Python llama.cpp engine (#1925) commit 6a1350c Author: Tushar Sharma <tusharma@nvidia.com> Date: Mon Jul 14 14:56:36 2025 -0700 build: minor improvements to sglang dockerfile (#1917) commit e2a619b Author: Neelay Shah <neelays@nvidia.com> Date: Mon Jul 14 14:52:53 2025 -0700 fix: remove environment variable passing (#1911) Signed-off-by: Neelay Shah <neelays@nvidia.com> Co-authored-by: Neelay Shah <neelays@a4u8g-0057.ipp2u2.colossus.nvidia.com> commit 3d17a49 Author: Schwinn Saereesitthipitak <17022745+galletas1712@users.noreply.github.com> Date: Mon Jul 14 14:41:56 2025 -0700 refactor: remove dynamo build (#1778) Signed-off-by: Schwinn Saereesitthipitak <17022745+galletas1712@users.noreply.github.com> commit 3e0cb07 Author: Anant Sharma <anants@nvidia.com> Date: Mon Jul 14 15:43:48 2025 -0400 fix: copy attributions and license to trtllm runtime container (#1916) commit fc36bf5 Author: ishandhanani <82981111+ishandhanani@users.noreply.github.com> Date: Mon Jul 14 12:31:49 2025 -0700 feat: receive kvmetrics from sglang scheduler (#1789) Co-authored-by: zixuanzhang226 <zixuanzhang@bytedance.com> commit df91fce Author: Yan Ru Pei <yanrpei@gmail.com> Date: Mon Jul 14 12:24:04 2025 -0700 feat: prefill aware routing (#1895) commit ad8ad66 Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 15:20:35 2025 -0400 feat: Shrink the ai-dynamo wheel by 35 MiB (#1918) Remove http and llmctl binaries. They have been unused for a while. commit 480b41d Author: Graham King <grahamk@nvidia.com> Date: Mon Jul 14 15:06:45 2025 -0400 feat: Python frontend / ingress node (#1912)
This was referenced Aug 25, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary by CodeRabbit