[Feature] Refactor /v1/abort_requests Endpoint

[qwes5s5]

Motivation

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick，PR标题需遵循格式，在最开始加上[Cherry-Pick]标签，以及最后面加上原PR ID，例如[Cherry-Pick][CI] Add check trigger and logic(#5191)

Currently, the /v1/abort_requests endpoint is synchronous, blocking the response until all target requests have been processed.

Modifications

Refactor this endpoint to an asynchronous model: it should return a response immediately upon receipt, offloading the actual interruption logic and result handling to the backend engine.

Usage or Command

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!

[PaddlePaddle-bot]

🤖 AI Code Review | 2026-04-24 20:48:11

📋 Review 摘要

PR 概述：将 /v1/abort_requests 端点从同步阻塞模式重构为异步 fire-and-forget 模式，同时提取 _resolve_abort_targets / _build_abort_results 两个辅助方法，删除 _wait_abort_complete。
变更范围：entrypoints/（api_server、engine_client、router）、engine/common_engine.py、scheduler/local_scheduler.py
影响面 Tag：APIServer Engine Scheduler

---

问题

级别	文件	概述
🔴 Bug	local_scheduler.py:136	ids_read_cursor 在删除末尾元素时越界
🔴 兼容性	api_server.py:500	/v1/abort_requests 响应格式破坏性变更
🟡 建议	engine_client.py:1142	abort_reqs() 无参调用时 engine 侧 KeyError 风险
🟡 建议	router.py:600	_forward_abort 无超时，挂起连接将永驻 _background_tasks

---

总体评价

重构方向正确，异步化可以避免接口阻塞；response_processors.py 中对空 token_ids 的防御性修复也是必要的。但 _recycle 的游标越界是真实的运行时崩溃风险，响应格式变更会静默破坏现有客户端，需要在合并前处理。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 75.00000% with 16 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@791484a). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/engine/common_engine.py	72.09%	10 Missing and 2 partials ⚠️
fastdeploy/entrypoints/engine_client.py	33.33%	2 Missing ⚠️
fastdeploy/scheduler/local_scheduler.py	50.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #7615   +/-   ##
==========================================
  Coverage           ?   71.66%           
==========================================
  Files              ?      419           
  Lines              ?    57804           
  Branches           ?     9064           
==========================================
  Hits               ?    41424           
  Misses             ?    13554           
  Partials           ?     2826           

Flag	Coverage Δ
GPU	71.66% <75.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.