feat(gateway): make vLLM sampler requests asynchronous and rate-limited#119
Closed
droot wants to merge 1 commit into
Closed
feat(gateway): make vLLM sampler requests asynchronous and rate-limited#119droot wants to merge 1 commit into
droot wants to merge 1 commit into
Conversation
Collaborator
Author
|
Pl. do not merge it. I haven't tested it yet. |
- Dispatch vLLM token generation requests to a background task instead of blocking the FastAPI handler, aligning its async behavior with the Torch sampler backend. - Introduce VLLM_CONCURRENCY_LIMIT (default 512) and _vllm_semaphore to prevent socket/file-descriptor exhaustion and connection drop errors under heavy surges. - Maintain a global _background_tasks set to hold strong references to running background tasks and prevent premature garbage collection.
droot
force-pushed
the
feature/async-vllm-sampling
branch
from
4f7d79e to
7c1a533
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Dispatch vLLM token generation requests to a background task instead of blocking the FastAPI handler, aligning its async behavior with the Torch sampler backend.
Introduce VLLM_CONCURRENCY_LIMIT (default 512) and _vllm_semaphore to prevent socket/file-descriptor exhaustion and connection drop errors under heavy surges.
Maintain a global _background_tasks set to hold strong references to running background tasks and prevent premature garbage collection.