Retry fallback after network errors and HTTP failures

docs/src/content/docs/en/configuration/routing.md

@@ -162,7 +162,7 @@ Network errors move to the next attempt. Status-code fallback depends on the mod
 | Retry | `408`, `409`, `429`, `5xx` |
 | Fallback targets | Any `4xx` or `5xx` |
 
-For `429` rate-limit responses, CCR waits before the next attempt. It honors `Retry-After` when the upstream provides it; otherwise it uses exponential backoff starting at 1 second and capped at 30 seconds per attempt.
+Before moving to the next attempt, CCR waits for every fallback-triggering failure, including network errors. It honors a positive `Retry-After` header when the upstream provides one; otherwise it uses exponential backoff starting at 1 second and capped at 30 seconds per attempt.
 
 **Fallback targets** also switches on `4xx` because model-not-found, auth, or provider-side rejection errors may only affect the current target. If the fallback model works, the request can still succeed.
 

docs/src/content/docs/zh/configuration/routing.md

@@ -162,7 +162,7 @@ Fallback 处理请求失败后的降级。第一次选模型由路由完成；
 | 继续重试 | `408`、`409`、`429`、`5xx` |
 | 失败降级目标 | 任意 `4xx` 或 `5xx` |
 
-对于 `429` 限流响应，CCR 会在下一次尝试前等待。上游提供 `Retry-After` 时会优先遵守；否则使用从 1 秒开始、单次最多 30 秒的指数退避。
+进入下一次尝试前，CCR 会对每个触发 Fallback 的失败进行等待，包括网络错误。上游提供正数 `Retry-After` 时会优先遵守；否则使用从 1 秒开始、单次最多 30 秒的指数退避。
 
 **失败降级目标** 对 `4xx` 也会切换，是因为模型不存在、鉴权或供应商侧拒绝等错误可能只影响当前目标。切换后如果备用模型可用，请求仍然可以成功。