The soak check measures `now - dev_HEAD_timestamp >= 24h`, which means
every new merge into dev resets the clock. Under any non-trivial merge
cadence, dev never satisfies the soak gate and master stagnates: PRs
#437/#438/#439/#440/#442 all stacked on dev for a week with no
promotion.

Add a `force_skip_soak` boolean input to workflow_dispatch (default
false). Schedule cron remains untouched and continues to enforce the
24h rule. Only manual triggers can bypass, and only when the operator
explicitly sets the input to true — intended for sync-back / merge
commits whose content has actually been on dev long enough but whose
HEAD timestamp is misleadingly fresh.

Test plan: yaml syntax verified via js-yaml. Functional verification
will be the next manual workflow_dispatch run with the input set.

Co-authored-by: icebear0828 <icebear0828@users.noreply.github.com>

…, proxy diag logs (#439)

* feat(dashboard): range hit-rate card + chart panel + 5min granularity, proxy diag logs

- UsageStats 新增「时段命中率」卡片（按当前时间窗口聚合 cached/input）
- UsageChart 新增「Hit Rate Over Time」面板：折线 + 数据点 dot，input=0 自动跳过
- 后端 + 前端支持 `five_min` granularity（5 分钟一桶，配合默认 5min snapshot）
- 新增 Last 1h / Last 6h 时间窗，granularity 切换时自动收敛兼容窗口
- shared/utils/usage-stats.ts 抽取 formatHitRate / sumWindow / formatUsageNumber，
  补单测覆盖边界（input=0、<0.01% 截断、windowed 求和）
- proxy-handler 入口/Usage 日志补诊断字段：rid / conv / key / prev=<src>:tail8 /
  tools=N / hit=X.X%，便于排查 prompt-cache 命中率为何偏低

* fix(proxy): send x-codex-installation-id + log resume reason for cache routing

- 上游请求新增 `x-codex-installation-id` header + body 内 `client_metadata`
  对齐真实 codex CLI（core/src/client.rs:874）。优先复用 `~/.codex/installation_id`，
  否则在 `data/installation_id` 持久化生成的 UUID。HTTP / WS / compact 三条路径都带。
- proxy-handler 入口日志加 `resume=on|off:<reason>` 字段，区分 implicit resume
  实际激活 vs 因 instr_diff / acct_mismatch / missing_tool_calls 等被否决的情况。
  `evaluateImplicitResume()` 替代 `shouldActivateImplicitResume()` 作为内部判定，
  保留旧 API 作为薄包装。
- 实测：installation_id 单独不能修复缓存命中率抖动（见后续 WS 池修复 PR）。

---------

Co-authored-by: icebear0828 <icebear0828@users.noreply.github.com>

The soak check measures `now - dev_HEAD_timestamp >= 24h`, which means
every new merge into dev resets the clock. Under any non-trivial merge
cadence, dev never satisfies the soak gate and master stagnates: PRs
#437/#438/#439/#440/#442 all stacked on dev for a week with no
promotion.

Add a `force_skip_soak` boolean input to workflow_dispatch (default
false). Schedule cron remains untouched and continues to enforce the
24h rule. Only manual triggers can bypass, and only when the operator
explicitly sets the input to true — intended for sync-back / merge
commits whose content has actually been on dev long enough but whose
HEAD timestamp is misleadingly fresh.

Test plan: yaml syntax verified via js-yaml. Functional verification
will be the next manual workflow_dispatch run with the input set.

Co-authored-by: icebear0828 <icebear0828@users.noreply.github.com>