[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"blog-strapi-article:zh:ai-api-retry-duplicate-request-cost-testing":3,"blog-article:\u002Fblog\u002Fai-api-retry-duplicate-request-cost-testing":1158,"blog-related:\u002Fblog\u002Fai-api-retry-duplicate-request-cost-testing":1159,"blog-translations:\u002Fblog\u002Fai-api-retry-duplicate-request-cost-testing":1160},{"available":4,"article":5,"relatedArticles":223,"translations":1010},true,{"source":6,"id":7,"documentId":8,"translationKey":8,"locale":9,"slug":10,"path":11,"_path":11,"title":12,"description":13,"excerpt":13,"image":14,"publishedAt":15,"updatedAt":16,"author":17,"tags":18,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":23,"bodyHtml":222},"strapi",3933,"g7sb9bj40ylxtdkkv9jp7u8b","zh","ai-api-retry-duplicate-request-cost-testing","\u002Fblog\u002Fai-api-retry-duplicate-request-cost-testing","AI API 重试与代理IP：如何确认没有重复请求和额外费用","验证 AI API 重试是否产生重复处理或额外费用，需要区分客户端发送尝试、服务端执行、实际用量和账单事件。文章介绍如何用受控靶子测试重试、按接口文档确认幂等能力，并用服务商用量明细核对结果。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_AI_API_cover_257ef04c34.webp","2026-11-04T06:49:08.009Z","2026-11-04T06:49:07.995Z","Global Proxies Team",[],false,0,3,"blocks",[24,30,36,53,57,61,65,69,83,87,91,95,99,103,107,125,129,133,137,141,155,159,171,175,179,183,187,191,198,202,208,212,218],{"type":25,"children":26},"paragraph",[27],{"text":28,"type":29},"重试会增加客户端发送尝试，但发送次数、服务端实际执行次数、计费用量和账单事件并不总是一一对应。失败请求是否计费取决于服务商、端点和实际处理状态，不能直接按请求次数乘单价推算。本文先用受控靶子验证客户端行为，再依据具体服务商文档和用量明细核对费用。","text",{"type":31,"level":32,"children":33},"heading",2,[34],{"text":35,"type":29},"哪些现象可以说明请求被重复发送",{"type":37,"format":38,"children":39},"list","unordered",[40,45,49],{"type":41,"children":42},"list-item",[43],{"text":44,"type":29},"信号一：靶子接收次数高于逻辑请求数，提示可能有重试、重定向或并发尝试，需按尝试 ID 继续核对。",{"type":41,"children":46},[47],{"text":48,"type":29},"信号二：相同幂等键出现多条访问记录，不能单凭访问次数判断执行次数；还要看处理事件和去重结果。",{"type":41,"children":50},[51],{"text":52,"type":29},"信号三：服务商用量明细与本地预期不符，作为待查信号，不能直接认定重复计费。",{"type":25,"children":54},[55],{"text":56,"type":29},"这些信号都只用于触发进一步核查。分别记录逻辑请求 ID、尝试 ID、服务商请求 ID、服务端处理事件、输入\u002F输出\u002F缓存用量及价格版本；只有证据链能对齐时，才判断是否重复执行或产生额外费用。",{"type":31,"level":32,"children":58},[59],{"text":60,"type":29},"为什么要用计数靶子替代真实付费接口",{"type":25,"children":62},[63],{"text":64,"type":29},"不要拿真实付费接口做故障注入。搭建只在测试环境可达的受控服务，记录逻辑 ID、尝试 ID、到达时间、测试请求标识、响应和服务端处理次数；只在自有测试数据上按需延迟响应或返回 429\u002F5xx。请求体哈希仅在自有系统、获得授权且确有必要时保存，并做好脱敏。靶子结果只能说明客户端行为，不能代表第三方服务的计费规则。",{"type":25,"children":66},[67],{"text":68,"type":29},"在自建靶子中，可按测试计划构造以下响应：",{"type":37,"format":38,"children":70},[71,75,79],{"type":41,"children":72},[73],{"text":74,"type":29},"可以按需返回 429 并带 `Retry-After`，观察客户端是否遵守。",{"type":41,"children":76},[77],{"text":78,"type":29},"可以在收到请求后延迟响应，触发客户端超时重试。",{"type":41,"children":80},[81],{"text":82,"type":29},"可以返回 307\u002F308，观察是否重放请求体。",{"type":25,"children":84},[85],{"text":86,"type":29},"靶子只服务测试流量，不接入业务链路；实验结束后关闭，避免误用。",{"type":31,"level":32,"children":88},[89],{"text":90,"type":29},"何时可以依赖幂等键",{"type":25,"children":92},[93],{"text":94,"type":29},"自建靶子可以实现并验证幂等键；调用第三方 AI API 时，只有服务商文档明确说明该端点支持对应幂等键、去重窗口和请求体规则，才能依赖此能力。",{"type":25,"children":96},[97],{"text":98,"type":29},"对自建靶子，可在逻辑请求开始时生成键并供所有重试复用；明确去重窗口及同键、不同请求体和并发请求的处理规则。第三方端点若不支持幂等，不要通过随意添加请求头假设服务端会去重；超时后先查任务状态，无法确认执行且操作非幂等时不要盲目重试。",{"type":25,"children":100},[101],{"text":102,"type":29},"对第三方接口，记录具体端点、文档版本、支持的键名、有效窗口及不同请求体的行为；没有明确文档时，将幂等能力标为未确认。",{"type":31,"level":32,"children":104},[105],{"text":106,"type":29},"如何逐项验证重试什么时候会触发",{"type":37,"format":38,"children":108},[109,113,117,121],{"type":41,"children":110},[111],{"text":112,"type":29},"场景：读超时后重试 ｜ 发送尝试数：按记录填写 ｜ 服务端处理数：按受控靶子事件填写 ｜ 去重命中：按靶子契约填写 ｜ 费用核验：第三方账单待核",{"type":41,"children":114},[115],{"text":116,"type":29},"场景：连接重置后重试 ｜ 发送尝试数：按记录填写 ｜ 服务端处理数：连接建立前后分别记 ｜ 去重命中：按接口能力填写 ｜ 费用核验：不从连接数推算费用",{"type":41,"children":118},[119],{"text":120,"type":29},"场景：429 与 Retry-After ｜ 发送尝试数：分别记录等待前后尝试 ｜ 服务端处理数：按服务端实际处理记录 ｜ 去重命中：按契约核对 ｜ 费用核验：等待用量明细结算",{"type":41,"children":122},[123],{"text":124,"type":29},"场景：307 \u002F 308 跟随 ｜ 发送尝试数：记录各跳请求 ｜ 服务端处理数：核对各跳是否执行业务 ｜ 去重命中：按端点契约填写 ｜ 费用核验：以服务商用量明细为准",{"type":25,"children":126},[127],{"text":128,"type":29},"发送尝试、服务端处理、去重结果和费用是四种不同计数。对 429 按 `Retry-After` 前后的请求及观察窗口记录预期；同键两次访问但仅执行一次可能符合契约。账单尚未结算或缺少服务端处理记录时，标注“证据不足\u002F账单未结算”，不要判定通过或重复计费。",{"type":25,"children":130},[131],{"text":132,"type":29},"客户端重试前应依据具体 API 契约确认可重试状态、幂等支持和 `Retry-After` 处理。",{"type":31,"level":32,"children":134},[135],{"text":136,"type":29},"307 和 308 重定向为什么可能重放请求",{"type":25,"children":138},[139],{"text":140,"type":29},"客户端若自动跟随 307\u002F308，通常会尝试保留方法和请求体；是否产生第二次业务处理取决于客户端策略、请求体可重放性以及原请求是否已执行。重定向由客户端按响应处理，换代理出口不决定业务请求应计一次还是两次。用自有靶子记录每一跳请求和处理事件，并注明客户端版本。",{"type":37,"format":38,"children":142},[143,147,151],{"type":41,"children":144},[145],{"text":146,"type":29},"客户端关闭自动跟随，或按端点契约逐跳处理并记录。",{"type":41,"children":148},[149],{"text":150,"type":29},"只有服务端明确支持幂等时才依赖其去重；否则先查询原任务状态。",{"type":41,"children":152},[153],{"text":154,"type":29},"记录重定向链路，确认中间有几跳。",{"type":25,"children":156},[157],{"text":158,"type":29},"浏览器跨源请求在符合 CORS 条件时可能出现 OPTIONS 预检；这不适用于所有客户端，不能把它一概视为 API 重复调用。",{"type":25,"children":160},[161,163,169],{"text":162,"type":29},"重定向语义可查 ",{"text":164,"type":29,"marks":165},"RFC 9110 HTTP Semantics",[166],{"href":167,"type":168},"https:\u002F\u002Fwww.rfc-editor.org\u002Frfc\u002Frfc9110","link",{"text":170,"type":29},"；具体客户端跟随行为以所用 SDK 文档和版本为准。",{"type":31,"level":32,"children":172},[173],{"text":174,"type":29},"如何对账请求次数、用量和实际费用",{"type":25,"children":176},[177],{"text":178,"type":29},"对账涉及客户端日志、靶子或网关计数、服务商用量面板。按逻辑请求、每次尝试、服务端处理事件、输入\u002F输出\u002F缓存用量和账单事件分别汇总。保存服务商请求 ID、价格版本、账单时间窗及结算状态；差额只作待查信号，按服务商用量明细和账单项核实，考虑结算延迟，不假定失败请求收费或按请求数乘单价计费。",{"type":31,"level":32,"children":180},[181],{"text":182,"type":29},"如何判定结果并加固重试策略",{"type":25,"children":184},[185],{"text":186,"type":29},"加固方向包括限制重试条件与次数、使用退避和随机抖动、按接口契约复用幂等键、分别记录逻辑请求与尝试 ID。结论可分为“符合约定”“发现重复执行\u002F费用”“证据不足\u002F账单未结算”；按事先约定的执行次数和费用标准验收，不能把所有重复发送一律判失败。",{"type":31,"level":32,"children":188},[189],{"text":190,"type":29},"FAQ",{"type":25,"children":192},[193],{"text":194,"type":29,"marks":195},"Q：幂等键由客户端生成，重试时复用就可以了吗？",[196],{"type":197},"bold",{"type":25,"children":199},[200],{"text":201,"type":29},"A：只在目标端点文档明确支持该键和去重规则时可以依赖；第三方不支持或规则不明时，不能仅靠客户端加头实现幂等。",{"type":25,"children":203},[204],{"text":205,"type":29,"marks":206},"Q：重试次数越多，成功率越高吗？",[207],{"type":197},{"type":25,"children":209},[210],{"text":211,"type":29},"A：不一定。次数增加会放大成本和排队压力，对非幂等请求还可能造成重复写入，应设上限并配合退避。",{"type":25,"children":213},[214],{"text":215,"type":29,"marks":216},"Q：怎么快速判断账单异常是不是重试造成的？",[217],{"type":197},{"type":25,"children":219},[220],{"text":221,"type":29},"A：先关联逻辑请求 ID、尝试 ID、服务商请求 ID和用量明细，再检查账单事件与结算状态。多条连接或访问记录不能单独证明重复执行或重复计费；普通 HTTPS 隧道通常也看不到请求体哈希。","",[224,466,737],{"source":6,"id":225,"documentId":226,"translationKey":226,"locale":9,"slug":227,"path":228,"_path":228,"title":229,"description":230,"excerpt":230,"image":231,"publishedAt":232,"updatedAt":233,"author":17,"tags":234,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":235,"bodyHtml":222},3937,"eg98712cpvp1nwo9plj025g4","websocket-reconnection-session-recovery-testing","\u002Fblog\u002Fwebsocket-reconnection-session-recovery-testing","WebSocket 断线重连：代理IP 出口切换后如何验证会话恢复","客户端显示已重连，用户却丢了消息或看到重复下单，是实时应用常见的假恢复。文章按客户端正常关闭、网络中断与服务端重启三类断线分别制造中断，比较恢复耗时、丢失与重复条数，再用消息序号、去重窗口和业务状态核对，确认恢复的是业务而不只是连接。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_Web_Socket_cover_2052233c88.webp","2026-11-04T06:50:15.056Z","2026-11-04T06:50:15.027Z",[],[236,240,244,254,258,262,266,270,274,285,289,311,315,319,333,337,341,345,363,367,371,375,379,383,387,391,405,409,413,423,433,436,442,446,452,456,462],{"type":25,"children":237},[238],{"text":239,"type":29},"日志里出现\"重连成功\"，并不代表业务恢复完成：消息可能缺失、重复处理或状态未同步。应按受控断线场景检查重连与业务状态，并根据协议契约验证恢复位置。本篇不预设断线一定发生在代理出口，也不把单一关闭码当作根因。",{"type":31,"level":32,"children":241},[242],{"text":243,"type":29},"怎样判断连接恢复是否已经变成业务恢复",{"type":37,"format":38,"children":245},[246,250],{"type":41,"children":247},[248],{"text":249,"type":29},"连接恢复：新的 WebSocket 完成握手，收到 101，可以收发消息。",{"type":41,"children":251},[252],{"text":253,"type":29},"业务恢复：断线期间产生的消息被补齐或明确丢弃，恢复后不重复、不乱序，业务状态与服务端一致。",{"type":25,"children":255},[256],{"text":257,"type":29},"连接恢复是必要条件，不是充分条件。客户端重连后如果没有带上恢复凭据，服务端可能按游标、时间窗或自身策略补发部分或全部消息，也可能直接拒绝恢复；策略不同，出问题的形态也不同：全量重放会重复，拒绝恢复会缺失。具体按哪种策略，以协议文档和测试结果为准。",{"type":31,"level":32,"children":259},[260],{"text":261,"type":29},"如何分别制造三种可控断线",{"type":25,"children":263},[264],{"text":265,"type":29},"不同断线方式会触发不同的关闭码与恢复路径，必须分开测：",{"type":25,"children":267},[268],{"text":269,"type":29},"1. 客户端主动关闭：调用 `close(1000)`，按应用预定策略验证停止或重新连接及状态处理；不自动重连不必然表示故障。\n2. 网络层中断：在自建测试网络中丢弃连接、发送 RST 或停止客户端进程等，记录故障点和方向。若客户端报告 1006，将其视为本地未收到有效 Close 帧的异常状态，不预设所有中断都会触发该码。\n3. 服务端主动断开：服务端关闭连接或重启，验证重连后的鉴权与恢复流程。",{"type":25,"children":271},[272],{"text":273,"type":29},"在自建测试环境中制造中断，不对正常业务链路注入故障。先做少量试跑检查测试流程，再根据业务 SLA、风险和观测波动确定正式覆盖次数。断线时机可覆盖空闲期、消息传输中和鉴权刚完成时，并记录 TCP RST、丢包、进程退出、心跳超时等实际证据。",{"type":25,"children":275},[276,278,283],{"text":277,"type":29},"网络层中断的定位可参考",{"text":279,"type":29,"marks":280},"代理 IP 故障排查中心",[281],{"href":282,"type":168},"https:\u002F\u002Fhelp.quanqiudaili.com\u002Fdocs\u002FSmart-check",{"text":284,"type":29},"。",{"type":31,"level":32,"children":286},[287],{"text":288,"type":29},"重连策略应该观察哪些指标",{"type":37,"format":38,"children":290},[291,295,299,303,307],{"type":41,"children":292},[293],{"text":294,"type":29},"首次重连延迟：从检测到断线到发出重连请求的时间。",{"type":41,"children":296},[297],{"text":298,"type":29},"重连间隔序列：是否指数退避，是否带随机抖动。",{"type":41,"children":300},[301],{"text":302,"type":29},"重连次数上限：达到上限后是放弃还是降级。",{"type":41,"children":304},[305],{"text":306,"type":29},"鉴权是否重做：令牌过期时是否重新获取。",{"type":41,"children":308},[309],{"text":310,"type":29},"是否携带恢复凭据：resume token、最后一条消息的序号或事件 ID。",{"type":25,"children":312},[313],{"text":314,"type":29},"退避没有抖动时，大量客户端会在同一时刻重连，形成二次冲击。验证时观察间隔序列是否递增且带随机量。",{"type":31,"level":32,"children":316},[317],{"text":318,"type":29},"常见的三种会话恢复机制怎么验证",{"type":37,"format":38,"children":320},[321,325,329],{"type":41,"children":322},[323],{"text":324,"type":29},"服务端保留游标：客户端带上最后确认的序号，服务端从下一条开始推送。",{"type":41,"children":326},[327],{"text":328,"type":29},"客户端拉取历史：重连后主动请求断线时间窗内的消息，再切回实时流。",{"type":41,"children":330},[331],{"text":332,"type":29},"无状态重放：服务端按时间窗重放，客户端负责去重。",{"type":25,"children":334},[335],{"text":336,"type":29},"序号的单调范围可能限于某个分区、订阅或过滤条件；先确认序号作用域、客户端确认点、乱序窗口和过滤规则。三种机制按产品契约分别验证：游标是否被接受、历史拉取与实时流是否衔接、重放窗口是否被正确去重。缺少客户端 token 不代表服务端没有已保存的恢复位置。",{"type":31,"level":32,"children":338},[339],{"text":340,"type":29},"如何检查补发消息是否重复或乱序",{"type":25,"children":342},[343],{"text":344,"type":29},"恢复阶段最容易出现\"补多了\"和\"补乱了\"。校验方法：",{"type":37,"format":38,"children":346},[347,351,355,359],{"type":41,"children":348},[349],{"text":350,"type":29},"每条业务消息带唯一 ID 与单调递增序号。",{"type":41,"children":352},[353],{"text":354,"type":29},"客户端记录已处理 ID 的窗口，重复 ID 直接丢弃。",{"type":41,"children":356},[357],{"text":358,"type":29},"收到消息时按已确认的序号作用域检查连续性；出现缺口后，按协议选择缓冲等待、补拉历史或进入降级\u002F重新同步，不规定所有系统都必须立即停止消费。",{"type":41,"children":360},[361],{"text":362,"type":29},"分别记录重复接收、去重后重复产生的业务副作用、最终缺失条数，作为恢复质量指标。",{"type":25,"children":364},[365],{"text":366,"type":29},"序号不能只靠时间戳，同一毫秒内的多条消息会无法排序。服务端生成序号时保证单调，客户端只做校验与去重。",{"type":31,"level":32,"children":368},[369],{"text":370,"type":29},"如何确认重连后业务状态仍然一致",{"type":25,"children":372},[373],{"text":374,"type":29},"连接层和消息层都正常，业务仍可能不一致。核对三项：",{"type":25,"children":376},[377],{"text":378,"type":29},"1. 关键状态字段：账户余额、订单状态、会话归属是否与服务端一致。\n2. 未确认操作：断线前发出但未收到确认的请求，重连后是否重发、是否被服务端去重。\n3. 幂等性：重发的写操作是否只生效一次（去重机制见上一节）。",{"type":25,"children":380},[381],{"text":382,"type":29},"业务核对要读服务端的权威状态，不能只信客户端本地缓存。本地缓存与服务端不一致时，以服务端为准并记录差异。",{"type":31,"level":32,"children":384},[385],{"text":386,"type":29},"如何记录恢复时间并完成验收",{"type":25,"children":388},[389],{"text":390,"type":29},"恢复时间要拆成两个数：断线到新连接可用，断线到业务补齐完成。前者反映重连策略，后者反映恢复机制。只报前者会低估用户实际等待时间。",{"type":37,"format":38,"children":392},[393,397,401],{"type":41,"children":394},[395],{"text":396,"type":29},"断线类型：客户端 close(1000) ｜ 恢复耗时：1.2 s ｜ 丢失条数：0 ｜ 重复条数：0 ｜ 业务状态一致：是 ｜ 证据：序号连续、状态字段一致",{"type":41,"children":398},[399],{"text":400,"type":29},"断线类型：网络中断（1006） ｜ 恢复耗时：3.5 s ｜ 丢失条数：0 ｜ 重复条数：2 ｜ 业务状态一致：是 ｜ 证据：去重窗口命中、补齐完成",{"type":41,"children":402},[403],{"text":404,"type":29},"断线类型：服务端重启 ｜ 恢复耗时：8.0 s ｜ 丢失条数：0 ｜ 重复条数：0 ｜ 业务状态一致：是 ｜ 证据：重新鉴权、游标续传",{"type":25,"children":406},[407],{"text":408,"type":29},"表中数值仅为示例，实际按测试记录填写；网络中断示例中的 1006 仅表示该次客户端本地观察到异常关闭状态，不代表确定的根因。重复接收不必然是缺陷，需进一步核对去重后的业务副作用。",{"type":25,"children":410},[411],{"text":412,"type":29},"验收表包含断线类型、恢复耗时、最终缺失条数、重复接收数、重复业务副作用、状态一致性和证据位置。覆盖次数与阈值按业务 SLA、风险和测量波动事先确定；3 次或 5 次只能作为试跑示例，不能当成通用充分门槛。正常 `close(1000)` 应按产品预定策略验证停止或恢复，不把不自动重连直接判错。任一关键证据缺失，结论应标注“证据不足”或“观察到重连成功”，不能写“恢复完成”。",{"type":25,"children":414},[415,417,422],{"text":416,"type":29},"测试工具的接入参数可对照",{"text":418,"type":29,"marks":419},"代理设置帮助",[420],{"href":421,"type":168},"https:\u002F\u002Fhelp.quanqiudaili.com\u002Fdocs\u002Fproxy-Settings",{"text":284,"type":29},{"type":25,"children":424},[425,427,432],{"text":426,"type":29},"WebSocket 关闭状态和 Close 帧语义见 ",{"text":428,"type":29,"marks":429},"RFC 6455",[430],{"href":431,"type":168},"https:\u002F\u002Fwww.rfc-editor.org\u002Frfc\u002Frfc6455",{"text":284,"type":29},{"type":31,"level":32,"children":434},[435],{"text":190,"type":29},{"type":25,"children":437},[438],{"text":439,"type":29,"marks":440},"Q：重连成功但消息少了一段，问题出在哪？",[441],{"type":197},{"type":25,"children":443},[444],{"text":445,"type":29},"A：先对照客户端、服务端和网络事件确定可观测到的故障位置，再按协议检查恢复位置。恢复状态可能来自客户端凭据，也可能由服务端保存；没有客户端序号不能单独证明服务端无法续传。",{"type":25,"children":447},[448],{"text":449,"type":29,"marks":450},"Q：close code 1006 一定能复现吗？",[451],{"type":197},{"type":25,"children":453},[454],{"text":455,"type":29},"A：不保证。1006 是本地报告的异常关闭状态，表示没有收到有效 Close 帧；不同客户端实现和故障方式下的观测结果可能不同，需记录具体故障注入和客户端事件。",{"type":25,"children":457},[458],{"text":459,"type":29,"marks":460},"Q：重复消息一定要在客户端去重吗？",[461],{"type":197},{"type":25,"children":463},[464],{"text":465,"type":29},"A：取决于协议契约。服务端游标续传、时间窗重放和客户端去重的责任不同，应分别检查收到重复与重复业务副作用，不能只统计客户端收到的消息数。",{"source":6,"id":467,"documentId":468,"translationKey":468,"locale":9,"slug":469,"path":470,"_path":470,"title":471,"description":472,"excerpt":472,"image":473,"publishedAt":474,"updatedAt":475,"author":17,"tags":476,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":477,"bodyHtml":222},3936,"wiumgl5vcc6qclqx4xe26751","websocket-proxy-test-report-metrics","\u002Fblog\u002Fwebsocket-proxy-test-report-metrics","WebSocket 代理IP 测试报告应该记录哪些指标","一份只能写「连接成功、延迟正常」的 WebSocket 测试报告，问题不在结论而在字段。把报告拆成环境、连接、业务三层，逐项给出字段、单位、采集方式与用途，配合握手分段、断连码分布、重连次数与 RTT 分位数，形成可直接落表、可复核的模板。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_Web_Socket_cover_bc45bc1e13.webp","2026-11-04T06:50:00.175Z","2026-11-04T06:50:00.160Z",[],[478,482,486,500,504,508,512,530,534,547,551,573,577,581,603,607,611,615,649,653,657,661,679,683,687,691,695,704,707,713,717,723,727,733],{"type":25,"children":479},[480],{"text":481,"type":29},"一份只能写\"连接成功、延迟正常\"的 WebSocket 报告，问题在于缺少可复核字段。本文按环境、连接、业务三层列出报告项，并注明采集工具与可观测边界。浏览器标准 WebSocket 事件通常不能直接提供全部握手分段、原始响应头或每一跳耗时；拿不到的字段应标为 N\u002FA 并说明原因。",{"type":31,"level":32,"children":483},[484],{"text":485,"type":29},"WebSocket 代理测试报告应该分哪三层记录",{"type":37,"format":38,"children":487},[488,492,496],{"type":41,"children":489},[490],{"text":491,"type":29},"环境层：这次测试在什么条件下发生。地区、出口 IP 与 ASN、代理协议、客户端版本、测试时间窗。",{"type":41,"children":493},[494],{"text":495,"type":29},"连接层：连接本身发生了什么。握手各段耗时、会话寿命、关闭码、异常关闭原因。",{"type":41,"children":497},[498],{"text":499,"type":29},"业务层：连接之上传了什么。消息计数、往返时延、丢失与重复、顺序连续性、恢复耗时。",{"type":25,"children":501},[502],{"text":503,"type":29},"三层分开写的好处是定位快：业务层异常先看连接层是否断过，连接层异常再看环境层是否出口漂移。混在一起写，读者只能看到\"整体不好\"。",{"type":31,"level":32,"children":505},[506],{"text":507,"type":29},"握手和建连阶段要记录哪些字段",{"type":25,"children":509},[510],{"text":511,"type":29},"握手要分四段记录，不能只写总和：",{"type":37,"format":38,"children":513},[514,518,522,526],{"type":41,"children":515},[516],{"text":517,"type":29},"DNS 解析耗时（ms）。",{"type":41,"children":519},[520],{"text":521,"type":29},"TCP 建连耗时（ms）。",{"type":41,"children":523},[524],{"text":525,"type":29},"TLS 握手耗时（ms，wss 才有）。",{"type":41,"children":527},[528],{"text":529,"type":29},"Upgrade 往返耗时（ms）：请求发出到收到 `101 Switching Protocols`。",{"type":25,"children":531},[532],{"text":533,"type":29},"同时记录校验结果：`Sec-WebSocket-Accept` 是否与服务端计算值一致、协商的子协议与扩展是否与预期相同、响应头是否被中间设备改写。握手失败要记录失败阶段与错误码，而不是只记\"连接失败\"。",{"type":25,"children":535},[536,538,541,543,546],{"text":537,"type":29},"握手前的代理参数可对照",{"text":418,"type":29,"marks":539},[540],{"href":421,"type":168},{"text":542,"type":29},"；Upgrade 握手字段定义见 ",{"text":428,"type":29,"marks":544},[545],{"href":431,"type":168},{"text":284,"type":29},{"type":31,"level":32,"children":548},[549],{"text":550,"type":29},"连接持续多久、为什么断开要怎么记录",{"type":37,"format":38,"children":552},[553,557,561,565,569],{"type":41,"children":554},[555],{"text":556,"type":29},"会话观察时长（s）：从收到 101 到关闭或观察窗口结束；窗口结束仍在线单独计数，不当作自然断开。",{"type":41,"children":558},[559],{"text":560,"type":29},"关闭发起方：只有两端事件或网络日志能佐证时才填写；无法确认则记未知。",{"type":41,"children":562},[563],{"text":564,"type":29},"本地关闭事件码与实际 Close 帧分开记录。1006 是本地报告的异常关闭状态，不是线上发送的 Close 帧；正常码也不能单独证明业务正常。",{"type":41,"children":566},[567],{"text":568,"type":29},"异常关闭率：明确分子、分母和观察窗口；握手失败、已关闭连接与窗口结束仍在线的连接分别统计。",{"type":41,"children":570},[571],{"text":572,"type":29},"心跳配置：间隔与超时阈值（s），以及心跳是否被中间设备吞掉。",{"type":25,"children":574},[575],{"text":576,"type":29},"关闭事件码分布可帮助筛选样本，但不能仅凭 code 判定发起方或根因。1006 表示本地未观察到有效 Close 帧；需结合客户端、服务端或网络侧日志继续定位。",{"type":31,"level":32,"children":578},[579],{"text":580,"type":29},"重连过程应该记录哪些指标",{"type":37,"format":38,"children":582},[583,587,591,595,599],{"type":41,"children":584},[585],{"text":586,"type":29},"首次重连延迟（ms\u002Fs）：从检测到断线到发起重连。",{"type":41,"children":588},[589],{"text":590,"type":29},"重连次数与退避序列：是否按指数退避，是否有抖动。",{"type":41,"children":592},[593],{"text":594,"type":29},"重连成功率：成功建立新连接的比例。",{"type":41,"children":596},[597],{"text":598,"type":29},"恢复耗时：从断线到业务可用的时间。",{"type":41,"children":600},[601],{"text":602,"type":29},"是否携带恢复凭据：如 resume token、last sequence。",{"type":25,"children":604},[605],{"text":606,"type":29},"重连次数多不代表质量差，可能是服务端主动轮转；但没有退避、密集重连会把故障放大。报告应同时给次数与间隔分布。",{"type":31,"level":32,"children":608},[609],{"text":610,"type":29},"如何汇总三层字段并补充消息细节",{"type":25,"children":612},[613],{"text":614,"type":29},"下表把三层字段汇总成可直接落表的形式。",{"type":37,"format":38,"children":616},[617,621,625,629,633,637,641,645],{"type":41,"children":618},[619],{"text":620,"type":29},"字段组：握手 ｜ 字段：DNS \u002F TCP \u002F TLS \u002F Upgrade ｜ 单位：ms ｜ 采集方式：受控网络日志、协议工具或服务端\u002F代理日志；浏览器标准事件通常不可提供全部分段 ｜ 用途：定位可观测范围内的建连耗时",{"type":41,"children":622},[623],{"text":624,"type":29},"字段组：连接 ｜ 字段：观察时长、结束状态、仍在线数 ｜ 单位：s \u002F 次 ｜ 采集方式：连接事件与观察窗口 ｜ 用途：区分自然断开与测试截止",{"type":41,"children":626},[627],{"text":628,"type":29},"字段组：连接 ｜ 字段：本地关闭事件码、Close 帧收发、可证实的发起方 ｜ 单位：次数 ｜ 采集方式：客户端事件及可用两端日志；不可观测时记 N\u002FA ｜ 用途：作为定位线索，不单独归因",{"type":41,"children":630},[631],{"text":632,"type":29},"字段组：重连 ｜ 字段：重连次数与间隔 ｜ 单位：次 \u002F s ｜ 采集方式：重连逻辑埋点 ｜ 用途：评估退避是否合理",{"type":41,"children":634},[635],{"text":636,"type":29},"字段组：消息 ｜ 字段：收发计数 ｜ 单位：条 ｜ 采集方式：消息回调 ｜ 用途：核对是否丢消息",{"type":41,"children":638},[639],{"text":640,"type":29},"字段组：消息 ｜ 字段：RTT p50 \u002F p95 ｜ 单位：ms ｜ 采集方式：发送与回执时间戳 ｜ 用途：评估体感延迟",{"type":41,"children":642},[643],{"text":644,"type":29},"字段组：消息 ｜ 字段：序号缺口与重复数 ｜ 单位：条 ｜ 采集方式：序号连续性校验 ｜ 用途：判断顺序与去重",{"type":41,"children":646},[647],{"text":648,"type":29},"字段组：代理IP ｜ 字段：类型、出口和会话分配 ｜ 单位：枚举 \u002F 地址 \u002F ID ｜ 采集方式：供应商控制台、分配记录及受控回显 ｜ 用途：说明本次测试路径，不能单独证明物理路由",{"type":25,"children":650},[651],{"text":652,"type":29},"表外还要补两项：消息大小分布与发送频率，否则 RTT 无法横向比较。大消息和小消息的往返时间差异明显，混在一起算分位数会失真。",{"type":31,"level":32,"children":654},[655],{"text":656,"type":29},"地区和代理链路还要记录哪些信息",{"type":25,"children":658},[659],{"text":660,"type":29},"代理IP字段记录实际分配的产品类型、出口和会话变化；无法确认的属性应标为未知。",{"type":37,"format":38,"children":662},[663,667,671,675],{"type":41,"children":664},[665],{"text":666,"type":29},"地区：客户端所在地区，与出口地区分开写。",{"type":41,"children":668},[669],{"text":670,"type":29},"出口 IP 与 ASN：每条连接记录一次，用于验证地区标签。",{"type":41,"children":672},[673],{"text":674,"type":29},"代理协议与会话类型：HTTP、SOCKS5，短会话或长会话；品类按实际分配写动态或静态住宅 IP、动态或静态机房 IP，不写笼统的\"住宅或机房\"，并保留出口 IP 与分配时间。",{"type":41,"children":676},[677],{"text":678,"type":29},"链路跳数：若代理有多跳，记录跳数与各跳耗时。",{"type":25,"children":680},[681],{"text":682,"type":29},"出口 IP 是测试记录的重要字段，但需与服务端回显、代理分配记录和 DNS\u002F调度信息结合；它本身不能证明物理路径或目标服务调度。",{"type":31,"level":32,"children":684},[685],{"text":686,"type":29},"报告模板应该怎样组织和统计",{"type":25,"children":688},[689],{"text":690,"type":29},"模板按\"环境 → 连接 → 业务\"顺序排列，每部分先给汇总表，再列异常样本。统计时报告 p50、p95 和样本数，单次最快值不能代表整体表现。\n失败与超时单独计数，不混入成功样本的分位数。\n按配置地区、出口和时间窗分组；跨窗口比较时同时标明环境差异。",{"type":25,"children":692},[693],{"text":694,"type":29},"异常样本保留 request id、连接 ID、出口 IP、关闭事件、时间戳及可用证据。报告结论写明客户端版本、服务端环境、测试时长、采集工具和不可观测字段；详细握手数据需受控网络日志、协议工具或服务端\u002F代理日志支持。",{"type":25,"children":696},[697,699,702],{"text":698,"type":29},"连接异常样本可按",{"text":279,"type":29,"marks":700},[701],{"href":282,"type":168},{"text":703,"type":29},"的字段顺序补充。",{"type":31,"level":32,"children":705},[706],{"text":190,"type":29},{"type":25,"children":708},[709],{"text":710,"type":29,"marks":711},"Q：报告里必须记录出口 IP 吗？",[712],{"type":197},{"type":25,"children":714},[715],{"text":716,"type":29},"A：建议记录，并同时记代理IP类型。它是验证地区标签与判断出口漂移的重要直接证据之一，但要与服务端回显、代理记录一起看，缺失后地区结论只能靠推断。",{"type":25,"children":718},[719],{"text":720,"type":29,"marks":721},"Q：close code 只记 1006 够不够？",[722],{"type":197},{"type":25,"children":724},[725],{"text":726,"type":29},"A：不够。关闭码是协议层线索，不能单独判断发起方或根因。应同时记录本地事件、实际 Close 帧收发和可取得的两端日志，并明确异常关闭率的分子、分母和仍在线连接的处理方式。",{"type":25,"children":728},[729],{"text":730,"type":29,"marks":731},"Q：时延只报平均值可以吗？",[732],{"type":197},{"type":25,"children":734},[735],{"text":736,"type":29},"A：不建议。平均值会掩盖尾部延迟，应同时报 p50 与 p95，并注明样本量。",{"source":6,"id":738,"documentId":739,"translationKey":739,"locale":9,"slug":740,"path":741,"_path":741,"title":742,"description":743,"excerpt":743,"image":744,"publishedAt":745,"updatedAt":746,"author":17,"tags":747,"featured":19,"sortOrder":20,"readingTime":748,"bodyType":22,"bodyBlocks":749,"bodyHtml":222},3935,"apg9xmkn8f0bozwgyxntuhtj","python-test-proxy-environment-variable-isolation","\u002Fblog\u002Fpython-test-proxy-environment-variable-isolation","Python 自动化测试：代理IP 环境变量污染如何验证","同一份测试在开发机能过、到 CI 就失败，常常是残留的代理变量在捣乱。文章列出这些变量及其读取顺序，用 fixture 固定环境、请求前断言、回显出口 IP 证明链路，再给出清理检查与故意失败自检。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_Python_cover_58829f8376.webp","2026-11-04T06:49:42.792Z","2026-11-04T06:49:42.779Z",[],4,[750,754,758,768,772,776,780,784,802,806,810,814,818,822,830,834,838,842,846,850,860,864,868,894,898,902,906,920,924,928,932,946,950,965,969,973,977,980,986,990,996,1000,1006],{"type":25,"children":751},[752],{"text":753,"type":29},"同一份测试，在开发机上通过、在 CI 上失败，或者反过来——十有八九是环境里的代理变量在起作用。requests、urllib、httpx 等常见实现通常会读取环境变量，具体行为随库和版本而变；CI 镜像又常预置 `HTTP_PROXY`。判断\"有没有被污染\"不能靠肉眼扫代码，要靠在测试内部固定环境并断言出口。这些变量指向的代理IP，往往和用例约定的不是同一条。NO_PROXY 规则本身的行为验证属于另一个话题，本篇讲的是测试框架里的环境隔离。",{"type":31,"level":32,"children":755},[756],{"text":757,"type":29},"代理环境变量污染通常会表现为什么",{"type":37,"format":38,"children":759},[760,764],{"type":41,"children":761},[762],{"text":763,"type":29},"意外走代理：本应直连的请求，被环境变量送进了代理，出口 IP 与预期不符。",{"type":41,"children":765},[766],{"text":767,"type":29},"意外绕过代理：本应走代理的请求，被 `NO_PROXY` 或空值放行，直接连到了目标。",{"type":25,"children":769},[770],{"text":771,"type":29},"两种表现都会让断言\"通过得莫名其妙\"。前者常出现在本地缓存或 mock 场景，后者常出现在 CI 里变量拼写错误或作用域不对的场景。判定污染不看代码写了什么，看实际请求从哪里出去。",{"type":31,"level":32,"children":773},[774],{"text":775,"type":29},"哪些代理环境变量会生效，优先级如何判断",{"type":25,"children":777},[778],{"text":779,"type":29},"变量不是设了就生效，读取顺序、大小写和库的默认值都会改变结果；变量只描述地址和协议，不描述代理IP的身份，身份得另做出口校验。把\"默认\"当成待实测的假设：按库和版本分别验证，并补一个故意设错变量的负向用例。",{"type":31,"level":21,"children":781},[782],{"text":783,"type":29},"为什么要同时检查大小写变量和 ALL_PROXY",{"type":37,"format":38,"children":785},[786,790,794,798],{"type":41,"children":787},[788],{"text":789,"type":29},"变量：`HTTP_PROXY` \u002F `http_proxy` ｜ 作用范围：http 请求 ｜ Python 客户端是否读取：requests、urllib 默认读取 ｜ 测试中如何处理：fixture 内显式设置或清空",{"type":41,"children":791},[792],{"text":793,"type":29},"变量：`HTTPS_PROXY` \u002F `https_proxy` ｜ 作用范围：https 请求 ｜ Python 客户端是否读取：同上 ｜ 测试中如何处理：同上",{"type":41,"children":795},[796],{"text":797,"type":29},"变量：`ALL_PROXY` \u002F `all_proxy` ｜ 作用范围：未单独指定协议时 ｜ Python 客户端是否读取：requests 读取 ｜ 测试中如何处理：容易遗漏，需一并清理",{"type":41,"children":799},[800],{"text":801,"type":29},"变量：`NO_PROXY` \u002F `no_proxy` ｜ 作用范围：排除列表 ｜ Python 客户端是否读取：requests、urllib 读取 ｜ 测试中如何处理：边界用例单独覆盖",{"type":25,"children":803},[804],{"text":805,"type":29},"`ALL_PROXY` 最容易被漏：未单独指定协议时它接管全部请求，清理时不能只清 HTTP_PROXY 与 HTTPS_PROXY。",{"type":31,"level":21,"children":807},[808],{"text":809,"type":29},"环境变量大小写相同时应该相信哪一个",{"type":25,"children":811},[812],{"text":813,"type":29},"同名不同大小写时，取值顺序可能相反。requests 通过 `urllib.request.getproxies()` 取值，常见实现小写优先，但版本和平台不同结果会变，也有库只认大写。测试里不要依赖，统一由 fixture 显式设定，并在请求前断言生效值，再用实际出口校验。",{"type":31,"level":21,"children":815},[816],{"text":817,"type":29},"为什么没写代理的 requests 仍可能走代理",{"type":25,"children":819},[820],{"text":821,"type":29},"Python 的 `requests` 常见版本默认 `trust_env=True`，会读取环境配置；该开关还会影响 `.netrc` 等其他来源。需要隔离环境代理时，优先为测试创建专用 `Session` 并设置 `trust_env=False`，再显式配置测试所需的认证与 CA。这样也会关闭该 Session 的其他环境来源，需按实际需求补配。不要把只对 `http`、`https` 设置 `None` 当成覆盖所有代理环境变量的通用方法；应按 requests 版本、目标地址和实际出口验证。",{"type":25,"children":823},[824,826,829],{"text":825,"type":29},"代理环境变量与请求库的关系可对照",{"text":418,"type":29,"marks":827},[828],{"href":421,"type":168},{"text":284,"type":29},{"type":31,"level":32,"children":831},[832],{"text":833,"type":29},"如何用 fixture 固定环境并提前断言",{"type":25,"children":835},[836],{"text":837,"type":29},"在测试层做三件事：",{"type":25,"children":839},[840],{"text":841,"type":29},"1. autouse fixture 在用例开始前清空所有代理变量（大小写各一份），用例结束后恢复。\n2. 需要代理的用例通过参数显式注入，不依赖外部环境。\n3. 请求发出前断言：`os.environ` 中的值、`session.proxies`、`session.trust_env` 与预期一致。",{"type":25,"children":843},[844],{"text":845,"type":29},"```python\nimport pytest",{"type":25,"children":847},[848],{"text":849,"type":29},"@pytest.fixture(autouse=True)\ndef clean_proxy_env(monkeypatch):\nfor key in (\"HTTP_PROXY\", \"HTTPS_PROXY\", \"ALL_PROXY\",\n\"NO_PROXY\", \"http_proxy\", \"https_proxy\", \"all_proxy\",\n\"no_proxy\"):\nmonkeypatch.delenv(key, raising=False)\n```",{"type":25,"children":851},[852,854,859],{"text":853,"type":29},"需要测试 `NO_PROXY` 时，在对应参数化用例中单独设置目标值。`monkeypatch` 会在用例结束时恢复原环境；可加一个检查，确认所有大小写变量恢复到用例前状态。详情见 ",{"text":855,"type":29,"marks":856},"pytest monkeypatch 文档",[857],{"href":858,"type":168},"https:\u002F\u002Fdocs.pytest.org\u002Fen\u002Fstable\u002Fhow-to\u002Fmonkeypatch.html",{"text":284,"type":29},{"type":31,"level":32,"children":861},[862],{"text":863,"type":29},"NO_PROXY 哪些边界情况必须单独测试",{"type":25,"children":865},[866],{"text":867,"type":29},"`NO_PROXY` 的匹配规则容易想当然，至少要覆盖：",{"type":37,"format":38,"children":869},[870,874,878,882,886,890],{"type":41,"children":871},[872],{"text":873,"type":29},"精确域名与子域名：`example.com` 是否覆盖 `api.example.com`。",{"type":41,"children":875},[876],{"text":877,"type":29},"端口写法：`example.com:443` 与不带端口的差异。",{"type":41,"children":879},[880],{"text":881,"type":29},"IP 与 CIDR：单 IP、网段写法是否被支持。",{"type":41,"children":883},[884],{"text":885,"type":29},"本机地址：`localhost`、`127.0.0.1`、`::1`。",{"type":41,"children":887},[888],{"text":889,"type":29},"通配：`*` 是否让所有请求直连。",{"type":41,"children":891},[892],{"text":893,"type":29},"空值：变量存在但为空字符串时的行为。",{"type":25,"children":895},[896],{"text":897,"type":29},"每个边界写成一个用例，断言\"是否走了代理\"，而不是\"变量是否被读取\"。变量被读取但匹配失败，同样属于污染。",{"type":31,"level":32,"children":899},[900],{"text":901,"type":29},"如何证明请求真的走了预期出口",{"type":25,"children":903},[904],{"text":905,"type":29},"环境变量只是配置，出口才是结果。回显请求必须由同一个被测客户端、协议和目标匹配规则发出；另用受控目标的访问日志或代理到达记录核对真正的业务请求。不同目标可能命中不同的 `NO_PROXY` 规则，单独请求另一个回显域名不能证明被测请求走了同一路径。直连出口应现场测量，并注明 NAT 或 CI 共享出口的影响：",{"type":37,"format":38,"children":907},[908,912,916],{"type":41,"children":909},[910],{"text":911,"type":29},"期望走代理：出口应属于代理侧地址段。",{"type":41,"children":913},[914],{"text":915,"type":29},"期望直连：出口应属于本机公网地址段。",{"type":41,"children":917},[918],{"text":919,"type":29},"期望绕过：目标应命中 `NO_PROXY`，出口为直连地址。",{"type":25,"children":921},[922],{"text":923,"type":29},"回显服务用自建或受控端点。分别覆盖预期走代理、命中绕过规则、故意设置错误配置和恢复后的请求，并记录目标、时间、客户端版本与代理到达证据。代理 URL 只记录脱敏后的协议、主机、端口和配置来源；删除 userinfo、令牌及敏感查询参数，避免 CI 日志泄露凭据。",{"type":31,"level":32,"children":925},[926],{"text":927,"type":29},"如何隔离开发机和 CI 继承的代理环境",{"type":25,"children":929},[930],{"text":931,"type":29},"本地 shell 里的代理变量会通过 `pytest` 进程继承进来，CI 的容器镜像也可能预置。隔离手段：",{"type":37,"format":38,"children":933},[934,938,942],{"type":41,"children":935},[936],{"text":937,"type":29},"在 `pytest.ini` 或 `conftest.py` 里统一清理，不依赖开发者本地配置。",{"type":41,"children":939},[940],{"text":941,"type":29},"CI 任务显式声明环境，不继承 runner 的全局变量。",{"type":41,"children":943},[944],{"text":945,"type":29},"需要真实代理的集成测试单独标记，默认不跑。",{"type":25,"children":947},[948],{"text":949,"type":29},"区分\"单元测试\"与\"集成测试\"的边界很重要：前者不应依赖任何外部出口，后者才需要真实代理。",{"type":25,"children":951},[952,954,959,961,964],{"text":953,"type":29},"requests 的环境设置行为见 ",{"text":955,"type":29,"marks":956},"Requests 高级用法文档",[957],{"href":958,"type":168},"https:\u002F\u002Frequests.readthedocs.io\u002Fen\u002Fstable\u002Fuser\u002Fadvanced\u002F",{"text":960,"type":29},"；测试环境变量恢复可参考 ",{"text":855,"type":29,"marks":962},[963],{"href":858,"type":168},{"text":284,"type":29},{"type":31,"level":32,"children":966},[967],{"text":968,"type":29},"如何通过清理检查和故意失败验证测试有效",{"type":25,"children":970},[971],{"text":972,"type":29},"最后一步是验证测试本身有效：把预期出口故意写错，测试必须失败。如果仍然通过，说明断言没有真正落到出口上，前面所有隔离都只是形式。",{"type":25,"children":974},[975],{"text":976,"type":29},"清理检查同样要做：用例结束后确认 `os.environ` 已恢复，没有把变量泄漏给后续用例。多个用例串行时，泄漏会表现为\"只有第一个用例正常\"。",{"type":31,"level":32,"children":978},[979],{"text":190,"type":29},{"type":25,"children":981},[982],{"text":983,"type":29,"marks":984},"Q：代码里没写代理，还会被环境变量影响吗？",[985],{"type":197},{"type":25,"children":987},[988],{"text":989,"type":29},"A：会。requests 等常见实现默认读取环境变量，`trust_env` 为 True 时自动生效；具体行为以实际版本为准，测试里要显式清理或关闭，并回显出口验证。",{"type":25,"children":991},[992],{"text":993,"type":29,"marks":994},"Q：`NO_PROXY` 写了域名，为什么子域名还是走了代理？",[995],{"type":197},{"type":25,"children":997},[998],{"text":999,"type":29},"A：不同实现对子域名的匹配规则不同，需要按库的实际行为写用例验证，不能假设后缀匹配一定生效。",{"type":25,"children":1001},[1002],{"text":1003,"type":29,"marks":1004},"Q：CI 上通过、本地失败，先查什么？",[1005],{"type":197},{"type":25,"children":1007},[1008],{"text":1009,"type":29},"A：先比对变量是否存在、脱敏后的协议\u002F主机\u002F端口和配置来源，再用同一被测请求的目标日志或代理到达记录确认路径。不要打印完整代理 URL，其中可能含账号、密码或令牌。",[1011],{"source":6,"id":7,"documentId":8,"translationKey":8,"locale":9,"slug":10,"path":11,"_path":11,"title":12,"description":13,"excerpt":13,"image":14,"publishedAt":15,"updatedAt":16,"author":17,"tags":1012,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":1013,"bodyHtml":222},[],[1014,1017,1020,1031,1034,1037,1040,1043,1054,1057,1060,1063,1066,1069,1072,1086,1089,1092,1095,1098,1109,1112,1119,1122,1125,1128,1131,1134,1139,1142,1147,1150,1155],{"type":25,"children":1015},[1016],{"text":28,"type":29},{"type":31,"level":32,"children":1018},[1019],{"text":35,"type":29},{"type":37,"format":38,"children":1021},[1022,1025,1028],{"type":41,"children":1023},[1024],{"text":44,"type":29},{"type":41,"children":1026},[1027],{"text":48,"type":29},{"type":41,"children":1029},[1030],{"text":52,"type":29},{"type":25,"children":1032},[1033],{"text":56,"type":29},{"type":31,"level":32,"children":1035},[1036],{"text":60,"type":29},{"type":25,"children":1038},[1039],{"text":64,"type":29},{"type":25,"children":1041},[1042],{"text":68,"type":29},{"type":37,"format":38,"children":1044},[1045,1048,1051],{"type":41,"children":1046},[1047],{"text":74,"type":29},{"type":41,"children":1049},[1050],{"text":78,"type":29},{"type":41,"children":1052},[1053],{"text":82,"type":29},{"type":25,"children":1055},[1056],{"text":86,"type":29},{"type":31,"level":32,"children":1058},[1059],{"text":90,"type":29},{"type":25,"children":1061},[1062],{"text":94,"type":29},{"type":25,"children":1064},[1065],{"text":98,"type":29},{"type":25,"children":1067},[1068],{"text":102,"type":29},{"type":31,"level":32,"children":1070},[1071],{"text":106,"type":29},{"type":37,"format":38,"children":1073},[1074,1077,1080,1083],{"type":41,"children":1075},[1076],{"text":112,"type":29},{"type":41,"children":1078},[1079],{"text":116,"type":29},{"type":41,"children":1081},[1082],{"text":120,"type":29},{"type":41,"children":1084},[1085],{"text":124,"type":29},{"type":25,"children":1087},[1088],{"text":128,"type":29},{"type":25,"children":1090},[1091],{"text":132,"type":29},{"type":31,"level":32,"children":1093},[1094],{"text":136,"type":29},{"type":25,"children":1096},[1097],{"text":140,"type":29},{"type":37,"format":38,"children":1099},[1100,1103,1106],{"type":41,"children":1101},[1102],{"text":146,"type":29},{"type":41,"children":1104},[1105],{"text":150,"type":29},{"type":41,"children":1107},[1108],{"text":154,"type":29},{"type":25,"children":1110},[1111],{"text":158,"type":29},{"type":25,"children":1113},[1114,1115,1118],{"text":162,"type":29},{"text":164,"type":29,"marks":1116},[1117],{"href":167,"type":168},{"text":170,"type":29},{"type":31,"level":32,"children":1120},[1121],{"text":174,"type":29},{"type":25,"children":1123},[1124],{"text":178,"type":29},{"type":31,"level":32,"children":1126},[1127],{"text":182,"type":29},{"type":25,"children":1129},[1130],{"text":186,"type":29},{"type":31,"level":32,"children":1132},[1133],{"text":190,"type":29},{"type":25,"children":1135},[1136],{"text":194,"type":29,"marks":1137},[1138],{"type":197},{"type":25,"children":1140},[1141],{"text":201,"type":29},{"type":25,"children":1143},[1144],{"text":205,"type":29,"marks":1145},[1146],{"type":197},{"type":25,"children":1148},[1149],{"text":211,"type":29},{"type":25,"children":1151},[1152],{"text":215,"type":29,"marks":1153},[1154],{"type":197},{"type":25,"children":1156},[1157],{"text":221,"type":29},null,[],[]]