QTom
72e5876c64
feat(gateway): Cache-Driven RPM Buffer
- buffer 公式从 baseRPM/5 改为 concurrency + maxSessions
保留 baseRPM/5 作为 floor 向后兼容
- 粘性路径 fallback 新增 [StickyCacheMiss] 结构化日志
reason: rpm_red / gate_check / session_limit / wait_queue_full / account_cleared
- session_limit 路径跳过 wait queue 重试(RegisterSession 拒绝无副作用)
- 典型配置 buffer 从 3 提升至 13,大幅减少高峰期 Prompt Cache Miss
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-31 13:24:22 +08:00
..
2026-03-30 08:13:49 +00:00
2026-03-27 14:33:05 +08:00
2026-03-31 13:24:22 +08:00
2026-03-27 14:33:05 +08:00
2026-03-19 19:03:13 +08:00
2026-01-11 23:00:31 +08:00
2026-03-07 15:17:16 +08:00
2026-02-06 07:41:23 +08:00
2026-03-17 20:46:12 +08:00
2026-03-21 01:21:48 +08:00
2026-02-28 15:01:20 +08:00