feat: 流式空闲超时、健康监控查询优化、限流桶内存上限与维护清理修复

Close #233

Co-authored-by: AAEE86 <ppk0227@hotmail.com>

- cli_monitor_mixin: 引入 STREAM_IDLE_TIMEOUT_SECONDS（可通过环境变量配置），
  流传输开始后若超出空闲窗口无新 chunk 则提前取消并返回 504，避免长时间挂起
- stream_context: 新增 managed_recorded_bodies 上下文管理器，确保 chunks 在
  telemetry 完成后及时释放；stream_telemetry 使用该接口统一管理 response body 构建
- health endpoint: 将状态聚合改为 GROUP BY 直接统计，事件列表按 api_format
  单独查询，避免单次 limit 拉取大量记录导致的遗漏与性能问题；同时过滤不活跃
  provider/endpoint，与公开健康接口保持一致
- endpoint health service: 修正时间线数据按 endpoint_id 而非 key_id 聚合
- token_bucket: 引入 max_buckets/bucket_expiry 上限与定时清理，防止内存无限增长；
  修复 refill_rate=0 时 get_reset_time 除零异常；新增 _is_unlimited_rate_limit 判断
- maintenance_scheduler: 调整清理顺序（先删整行再按窗口清理），新增 newer_than
  边界参数，避免同一行在同一轮中被重复改写
- sync_execute: 新增 create_pending_usage 开关，允许已预创建记录的调用方跳过重复创建
- quota_reader / provider_ops balance: 小幅修复与健壮性提升
- Dockerfile: 添加 MALLOC_ARENA_MAX=2 环境变量以降低 gunicorn worker RSS
- 补充相关测试覆盖

tests/api/handlers/base/test_stream_context.py

@@ -1,3 +1,9 @@
+import os
+
+import pytest
+
+os.environ.setdefault("JWT_SECRET_KEY", "test-secret-key")
+
 from src.api.handlers.base import stream_context
 from src.api.handlers.base.stream_context import StreamContext
 
@@ -50,7 +56,36 @@ def test_reset_for_retry_clears_state() -> None:
     assert ctx.error_message is None
 
 
-def test_record_first_byte_time(monkeypatch) -> None:
+def test_release_recorded_chunks_clears_both_chunk_lists() -> None:
+    ctx = StreamContext(model="test-model", api_format="openai:chat")
+    ctx.parsed_chunks.append({"type": "client"})
+    ctx.provider_parsed_chunks.append({"type": "provider"})
+
+    ctx.release_recorded_chunks()
+
+    assert ctx.parsed_chunks == []
+    assert ctx.provider_parsed_chunks == []
+
+
+def test_managed_recorded_bodies_builds_then_releases_chunks() -> None:
+    ctx = StreamContext(model="test-model", api_format="openai:chat")
+    ctx.parsed_chunks.append({"type": "client"})
+    ctx.provider_parsed_chunks.append({"type": "provider"})
+    ctx.data_count = 1
+
+    with ctx.managed_recorded_bodies(123) as recorded_bodies:
+        assert recorded_bodies.response_body is not None
+        assert recorded_bodies.response_body["chunks"] == [{"type": "provider"}]
+        assert recorded_bodies.client_response_body is not None
+        assert recorded_bodies.client_response_body["chunks"] == [{"type": "client"}]
+
+    assert ctx.parsed_chunks == []
+    assert ctx.provider_parsed_chunks == []
+    assert recorded_bodies.response_body is None
+    assert recorded_bodies.client_response_body is None
+
+
+def test_record_first_byte_time(monkeypatch: pytest.MonkeyPatch) -> None:
     """测试记录首字时间"""
     ctx = StreamContext(model="claude-3", api_format="claude_messages")
     start_time = 100.0
@@ -63,7 +98,7 @@ def test_record_first_byte_time(monkeypatch) -> None:
     assert ctx.first_byte_time_ms == 12
 
 
-def test_record_first_byte_time_idempotent(monkeypatch) -> None:
+def test_record_first_byte_time_idempotent(monkeypatch: pytest.MonkeyPatch) -> None:
     """测试首字时间只记录一次"""
     ctx = StreamContext(model="claude-3", api_format="claude_messages")
     start_time = 100.0
@@ -82,7 +117,7 @@ def test_record_first_byte_time_idempotent(monkeypatch) -> None:
     assert first_value == second_value
 
 
-def test_reset_for_retry_clears_first_byte_time(monkeypatch) -> None:
+def test_reset_for_retry_clears_first_byte_time(monkeypatch: pytest.MonkeyPatch) -> None:
     """测试重试时清除首字时间"""
     ctx = StreamContext(model="claude-3", api_format="claude_messages")
     start_time = 100.0