mirror of
https://github.com/QuantumNous/new-api.git
synced 2026-05-06 09:31:36 +00:00
feat: guard new 504/524 status remaps with risk confirmation
This commit is contained in:
@@ -1949,6 +1949,19 @@
|
||||
"自动重试状态码": "Auto-retry status codes",
|
||||
"自动重试状态码格式不正确": "Invalid auto-retry status code format",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔": "Supports single status codes or inclusive ranges; separate with commas",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔;504 和 524 始终不重试,不受此处配置影响": "Supports single status codes or inclusive ranges; separate with commas. 504 and 524 are never retried and are not affected by this setting",
|
||||
"高危操作确认": "High-risk operation confirmation",
|
||||
"检测到以下高危状态码重定向规则": "Detected high-risk status-code redirect rules",
|
||||
"操作确认": "Operation confirmation",
|
||||
"我确认开启高危重试": "I confirm enabling high-risk retry",
|
||||
"高危状态码重试风险告知与免责声明Markdown": "### ⚠️ High-Risk Operation: Risk Notice and Disclaimer for 504/524 Retry\n\n[Background]\nBy default, this project does not retry for status codes `400` (bad request), `504` (gateway timeout), and `524` (timeout occurred). In many cases, `504` and `524` mean the request has reached the upstream AI service and processing has started, but the connection was closed due to long processing time.\n\nEnabling redirection/retry for these timeout status codes is a **high-risk operation**. Before enabling it, you must read and understand the consequences below:\n\n#### 1. Core Risks (Read Carefully)\n1. 💸 Duplicate/multiple billing risk: Most upstream AI providers **still charge** for requests that started processing but got interrupted by network timeout (`504`/`524`). If retry is triggered, a new upstream request will be sent, which can lead to **duplicate or multiple charges**.\n2. ⏳ Severe client timeout: If a single request already timed out, adding retries can multiply total latency and cause severe or unacceptable timeout behavior for your final client/caller.\n3. 💥 Request backlog and system crash risk: Forcing retries on timeout requests keeps threads and connections occupied for longer. Under high concurrency, this can cause serious backlog, exhaust system resources, trigger a cascading failure, and crash your proxy service.\n\n#### 2. Risk Acknowledgement\nIf you still choose to enable this feature, you acknowledge all of the following:\n\n- [ ] I have fully read and understood the risks and fully understand the destructive consequences of forcing retries for status codes `504` and `524`.\n- [ ] I have communicated with the upstream provider and confirmed that the timeout issue is an upstream bottleneck and cannot be resolved upstream at this time.\n- [ ] I voluntarily accept all duplicate/multiple billing risks and will not file issues or complaints in this project repository regarding billing anomalies caused by this retry behavior.\n- [ ] I voluntarily accept system stability risks, including severe client timeout and possible service crash. Any consequences caused by enabling this feature are my own responsibility.\n\n> **[Operation Confirmation]**\n> To unlock this feature, manually type the text below in the input box:\n> I understand the duplicate billing and crash risks, and confirm enabling it.",
|
||||
"高危状态码重试风险确认输入文本": "I understand the duplicate billing and crash risks, and confirm enabling it.",
|
||||
"高危状态码重试风险确认项1": "I have fully read and understood the risks and fully understand the destructive consequences of forcing retries for status codes 504 and 524.",
|
||||
"高危状态码重试风险确认项2": "I have communicated with the upstream provider and confirmed that the timeout issue is an upstream bottleneck and cannot be resolved upstream at this time.",
|
||||
"高危状态码重试风险确认项3": "I voluntarily accept all duplicate/multiple billing risks and will not file issues or complaints in this project repository regarding billing anomalies caused by this retry behavior.",
|
||||
"高危状态码重试风险确认项4": "I voluntarily accept system stability risks, including severe client timeout and possible service crash. Any consequences caused by enabling this feature are my own responsibility.",
|
||||
"高危状态码重试风险输入框占位文案": "Please type the exact text above",
|
||||
"高危状态码重试风险输入不匹配提示": "The input does not match the required text",
|
||||
"例如:401, 403, 429, 500-599": "e.g. 401,403,429,500-599",
|
||||
"自动选择": "Auto Select",
|
||||
"自定义充值数量选项": "Custom Recharge Amount Options",
|
||||
|
||||
@@ -1936,6 +1936,19 @@
|
||||
"自动重试状态码": "自动重试状态码",
|
||||
"自动重试状态码格式不正确": "自动重试状态码格式不正确",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔": "支持填写单个状态码或范围(含首尾),使用逗号分隔",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔;504 和 524 始终不重试,不受此处配置影响": "支持填写单个状态码或范围(含首尾),使用逗号分隔;504 和 524 始终不重试,不受此处配置影响",
|
||||
"高危操作确认": "高危操作确认",
|
||||
"检测到以下高危状态码重定向规则": "检测到以下高危状态码重定向规则",
|
||||
"操作确认": "操作确认",
|
||||
"我确认开启高危重试": "我确认开启高危重试",
|
||||
"高危状态码重试风险告知与免责声明Markdown": "### ⚠️ 高危操作:504/524 状态码重试风险告知与免责声明\n\n【背景提示】\n本项目默认对 `400`(请求错误)、`504`(网关超时)和 `524`(发生超时)状态码不进行重试。504 和 524 错误通常意味着**请求已成功送达上游 AI 服务,且上游正在处理,但因处理时间过长导致连接断开**。\n\n开启对此类超时状态码的重定向/重试属于**极高风险操作**。作为本开源项目的使用者,在开启该功能前,您必须仔细阅读并知悉以下严重后果:\n\n#### 一、 核心风险告知(请仔细阅读)\n1. 💸 双重/多重计费风险: 绝大多数 AI 上游厂商对于已经开始处理但因网络原因中断(504/524)的请求**依然会进行扣费**。此时若触发重试,将会向上游发起全新请求,导致您被**双重甚至多重计费**。\n2. ⏳ 客户端严重超时: 单次请求已经触发超时,叠加重试机制将会使总请求耗时成倍增加,导致您的最终客户端(或调用方)出现严重甚至完全无法接受的超时现象。\n3. 💥 请求积压与系统崩溃风险: 强制重试超时请求会长时间占用系统线程和连接数。在高并发场景下,这会导致严重的**请求积压**,进而耗尽系统资源,引发雪崩效应,导致您的整个代理服务崩溃。\n\n#### 二、 风险确认声明\n如果您坚持开启该功能,即代表您作出以下确认:\n\n- [ ] 我已充分阅读并理解**:本人已完整阅读上述全部风险提示,完全理解强制重试 `504` 和 `524` 状态码可能带来的破坏性后果。\n- [ ] **我已与上游沟通并确认**:本人确认,当前出现的超时问题属于上游服务的瓶颈。**本人已与上游提供商进行过沟通,确认上游无法解决该超时问题**,因此才采取强制重试方案作为妥协手段。\n- [ ] **我自愿承担计费损失**:本人知晓并接受由此产生的全部双重/多重计费风险,**承诺不会因重试导致的账单异常在本项目仓库中提交 Issue 或抱怨**。\n- [ ] **我自愿承担系统稳定性风险**:本人知晓该操作可能导致客户端严重超时及服务崩溃。若因本人开启此功能导致请求积压或服务不可用,后果由本人自行承担。\n\n> **【操作确认】\n> 为确认您已清晰了解上述风险,请在下方输入框内手动输入以下文字以解锁功能:\n> 我已了解多重计费与崩溃风险,确认开启",
|
||||
"高危状态码重试风险确认输入文本": "我已了解多重计费与崩溃风险,确认开启",
|
||||
"高危状态码重试风险确认项1": "我已充分阅读并理解:本人已完整阅读上述全部风险提示,完全理解强制重试 504 和 524 状态码可能带来的破坏性后果。",
|
||||
"高危状态码重试风险确认项2": "我已与上游沟通并确认:本人确认,当前出现的超时问题属于上游服务的瓶颈。本人已与上游提供商进行过沟通,确认上游无法解决该超时问题,因此才采取强制重试方案作为妥协手段。",
|
||||
"高危状态码重试风险确认项3": "我自愿承担计费损失:本人知晓并接受由此产生的全部双重/多重计费风险,承诺不会因重试导致的账单异常在本项目仓库中提交 Issue 或抱怨。",
|
||||
"高危状态码重试风险确认项4": "我自愿承担系统稳定性风险:本人知晓该操作可能导致客户端严重超时及服务崩溃。若因本人开启此功能导致请求积压或服务不可用,后果由本人自行承担。",
|
||||
"高危状态码重试风险输入框占位文案": "请完整输入上方文字",
|
||||
"高危状态码重试风险输入不匹配提示": "输入内容与要求不一致",
|
||||
"例如:401, 403, 429, 500-599": "例如:401,403,429,500-599",
|
||||
"自动选择": "自动选择",
|
||||
"自定义充值数量选项": "自定义充值数量选项",
|
||||
|
||||
@@ -1942,6 +1942,19 @@
|
||||
"自动重试状态码": "自動重試狀態碼",
|
||||
"自动重试状态码格式不正确": "自動重試狀態碼格式不正確",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔": "支援填寫單個狀態碼或範圍(含首尾),使用逗號分隔",
|
||||
"支持填写单个状态码或范围(含首尾),使用逗号分隔;504 和 524 始终不重试,不受此处配置影响": "支援填寫單個狀態碼或範圍(含首尾),使用逗號分隔;504 和 524 一律不重試,不受此處設定影響",
|
||||
"高危操作确认": "高風險操作確認",
|
||||
"检测到以下高危状态码重定向规则": "檢測到以下高風險狀態碼重定向規則",
|
||||
"操作确认": "操作確認",
|
||||
"我确认开启高危重试": "我確認開啟高風險重試",
|
||||
"高危状态码重试风险告知与免责声明Markdown": "### ⚠️ 高風險操作:504/524 狀態碼重試風險告知與免責聲明\n\n【背景提示】\n本專案預設對 `400`(請求錯誤)、`504`(閘道逾時)與 `524`(發生逾時)狀態碼不進行重試。504 與 524 錯誤通常代表**請求已成功送達上游 AI 服務,且上游正在處理,但因處理時間過長導致連線中斷**。\n\n開啟此類逾時狀態碼的重定向/重試屬於**極高風險操作**。作為本開源專案使用者,在開啟該功能前,您必須仔細閱讀並知悉以下嚴重後果:\n\n#### 一、 核心風險告知(請仔細閱讀)\n1. 💸 雙重/多重計費風險:多數 AI 上游廠商對於已開始處理但因網路原因中斷(504/524)的請求**仍然會扣費**。此時若觸發重試,將會向上游發起全新請求,導致您被**雙重甚至多重計費**。\n2. ⏳ 用戶端嚴重逾時:單次請求已觸發逾時,疊加重試機制會使總請求耗時成倍增加,導致最終用戶端(或呼叫方)出現嚴重甚至無法接受的逾時現象。\n3. 💥 請求積壓與系統崩潰風險:強制重試逾時請求會長時間占用系統執行緒與連線數。在高併發場景下,這將導致嚴重**請求積壓**,進而耗盡系統資源,引發雪崩效應,造成整個代理服務崩潰。\n\n#### 二、 風險確認聲明\n若您堅持開啟該功能,即代表您作出以下確認:\n\n- [ ] 我已充分閱讀並理解:本人已完整閱讀上述全部風險提示,完全理解強制重試 `504` 與 `524` 狀態碼可能帶來的破壞性後果。\n- [ ] 我已與上游溝通並確認:本人確認,當前逾時問題屬於上游服務瓶頸。本人已與上游供應商溝通,確認上游無法解決該逾時問題,因此才採取強制重試方案作為妥協手段。\n- [ ] 我自願承擔計費損失:本人知悉並接受由此產生的全部雙重/多重計費風險,承諾不會因重試導致的帳單異常在本專案倉庫提交 Issue 或抱怨。\n- [ ] 我自願承擔系統穩定性風險:本人知悉該操作可能導致用戶端嚴重逾時及服務崩潰。若因本人開啟此功能導致請求積壓或服務不可用,後果由本人自行承擔。\n\n> **【操作確認】**\n> 為確認您已清楚了解上述風險,請在下方輸入框內手動輸入以下文字以解鎖功能:\n> 我已了解多重計費與崩潰風險,確認開啟",
|
||||
"高危状态码重试风险确认输入文本": "我已了解多重計費與崩潰風險,確認開啟",
|
||||
"高危状态码重试风险确认项1": "我已充分閱讀並理解:本人已完整閱讀上述全部風險提示,完全理解強制重試 504 與 524 狀態碼可能帶來的破壞性後果。",
|
||||
"高危状态码重试风险确认项2": "我已與上游溝通並確認:本人確認,當前逾時問題屬於上游服務瓶頸。本人已與上游供應商溝通,確認上游無法解決該逾時問題,因此才採取強制重試方案作為妥協手段。",
|
||||
"高危状态码重试风险确认项3": "我自願承擔計費損失:本人知悉並接受由此產生的全部雙重/多重計費風險,承諾不會因重試導致的帳單異常在本專案倉庫提交 Issue 或抱怨。",
|
||||
"高危状态码重试风险确认项4": "我自願承擔系統穩定性風險:本人知悉該操作可能導致用戶端嚴重逾時及服務崩潰。若因本人開啟此功能導致請求積壓或服務不可用,後果由本人自行承擔。",
|
||||
"高危状态码重试风险输入框占位文案": "請完整輸入上方文字",
|
||||
"高危状态码重试风险输入不匹配提示": "輸入內容與要求不一致",
|
||||
"例如:401, 403, 429, 500-599": "例如:401,403,429,500-599",
|
||||
"自动选择": "自動選擇",
|
||||
"自定义充值数量选项": "自訂儲值數量選項",
|
||||
|
||||
Reference in New Issue
Block a user