[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"blog-strapi-article:zh:playwright-context-proxy-egress-validation":3,"blog-article:\u002Fblog\u002Fplaywright-context-proxy-egress-validation":610,"blog-related:\u002Fblog\u002Fplaywright-context-proxy-egress-validation":611,"blog-translations:\u002Fblog\u002Fplaywright-context-proxy-egress-validation":612},{"available":4,"article":5,"relatedArticles":147,"translations":523},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":146},"strapi",3981,"ovyz1soa19lqmiqf0xtdsqsh","zh","playwright-context-proxy-egress-validation","\u002Fblog\u002Fplaywright-context-proxy-egress-validation","Playwright 自动化如何验证每个测试上下文使用了预期的代理 IP 出口","在每个测试上下文内校验出口 IP 与 ASN，记录代理参数、断言和失败表现；复用连接或并行运行时，先确认出口没有串到其他用例。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_Playwright_cover_e5b636a56e.webp","2026-11-06T07:38:47.723Z","2026-11-06T07:38:47.707Z","Global Proxies Team",[],false,0,2,"blocks",[24,30,35,39,43,47,51,63,67,71,82,86,90,94,98,102,106,125,129,134,138,142],{"type":25,"children":26},"paragraph",[27],{"text":28,"type":29},"要验证 Playwright 的每个 context 是否使用了预期出口，先分清两级代理配置：浏览器启动时设置的是默认代理，每个 context 也可以单独配置。设置完成后，在 context 创建后立即检查出口 IP；并发测试时，还要确认每个用例拿到的是自己的预期出口。","text",{"type":31,"level":21,"children":32},"heading",[33],{"text":34,"type":29},"Playwright 的代理配置分为浏览器级和 context 级",{"type":25,"children":36},[37],{"text":38,"type":29},"`browserType.launch({ proxy })` 会作用到所有 context；要每个 context 不同出口，必须配在 `browser.newContext({ proxy })` 上。这是 Playwright 的 context 级选项，具体可用参数和浏览器支持范围以所装版本的官方文档为准（`playwright.dev\u002Fdocs\u002Fapi\u002Fclass-browser#browser-new-context`）。",{"type":25,"children":40},[41],{"text":42,"type":29},"```js\nconst context = await browser.newContext({\nproxy: { server: \"\u003C代理IP地址和端口>\", username, password }\n});\n```",{"type":25,"children":44},[45],{"text":46,"type":29},"如果只在浏览器启动时设置代理，多个 context 可能会共用同一个出口。测试仍可能运行成功，因此应在每个 context 中请求回显 IP，并与该用例的预期出口逐一比对。",{"type":31,"level":21,"children":48},[49],{"text":50,"type":29},"为每个 context 设置代理并验证实际出口",{"type":25,"children":52},[53,55,61],{"text":54,"type":29},"context 创建后先请求回显 IP 的接口，比对“预期出口”和“实际出口”，断言通过再跑业务用例。断言放在业务之前，能更早暴露问题，但不能仅凭它定责：回显不符时先标“路径或分配待查”，再分段排查配置与运行环境。配置 context 代理时，可对照",{"text":56,"type":29,"marks":57},"代理设置帮助",[58],{"href":59,"type":60},"https:\u002F\u002Fhelp.quanqiudaili.com\u002Fdocs\u002Fproxy-Settings","link",{"text":62,"type":29},"核对协议与认证参数。",{"type":25,"children":64},[65],{"text":66,"type":29},"回显接口应使用自有或已获授权的稳定端点。固定出口时，将响应中的 IP 与分配值比较；动态出口则根据产品的分配规则验证本次结果，不能一概要求每次都返回同一个 IP。页面请求和 API 请求应分别记录用例、context、时间和回显结果，便于与代理侧的分配记录核对。日志中不要保存用户名或密码。",{"type":31,"level":21,"children":68},[69],{"text":70,"type":29},"分别验证页面请求和 API 请求的代理出口",{"type":25,"children":72},[73,75,80],{"text":74,"type":29},"页面请求通过 `page` 发出，API 请求则可使用 `context.request`（Playwright 的 APIRequestContext）。分别检查两类请求的实际出口，避免只验证网页流量。配置时若不确定主机、端口或认证字段，可对照",{"text":76,"type":29,"marks":77},"代理参数说明",[78],{"href":79,"type":60},"https:\u002F\u002Fhelp.quanqiudaili.com\u002Fdocs\u002Fintro",{"text":81,"type":29},"。",{"type":31,"level":21,"children":83},[84],{"text":85,"type":29},"Playwright 代理配置常见问题及处理方法",{"type":25,"children":87},[88],{"text":89,"type":29},"如果所有 context 都显示同一个出口，先检查是否在 `launch` 层设置了全局代理。出现 407 或连接被拒时，核对用户名和密码是否传入。某个浏览器不接受该配置时，按所用 Playwright 版本的官方说明确认支持范围。复用 `storageState` 后出口断言不符合预期，则分别检查代理配置和会话状态。",{"type":31,"level":21,"children":91},[92],{"text":93,"type":29},"将出口断言加入 fixture 并记录验证结果",{"type":25,"children":95},[96],{"text":97,"type":29},"把出口断言放进 fixture（测试用例共享的初始化逻辑），断言失败就让用例失败，而不是只记录日志。并发运行时，确保每个 context 与对应的预期出口一一匹配。",{"type":25,"children":99},[100],{"text":101,"type":29},"每条测试记录应包含 context 名称、预期和实际出口、断言结果，以及浏览器和 Playwright 版本。测试仅在自有或已获授权的环境中运行；若出口不符，先停止该用例，再检查代理配置层级。",{"type":25,"children":103},[104],{"text":105,"type":29},"配置选项以官方文档为准：",{"type":107,"format":108,"children":109},"list","unordered",[110,118],{"type":111,"children":112},"list-item",[113],{"text":114,"type":29,"marks":115},"Playwright 网络与代理配置",[116],{"href":117,"type":60},"https:\u002F\u002Fplaywright.dev\u002Fdocs\u002Fnetwork",{"type":111,"children":119},[120],{"text":121,"type":29,"marks":122},"`browser.newContext()` 的 `proxy` 选项说明",[123],{"href":124,"type":60},"https:\u002F\u002Fplaywright.dev\u002Fdocs\u002Fapi\u002Fclass-browser",{"type":31,"level":21,"children":126},[127],{"text":128,"type":29},"常见问题",{"type":31,"level":130,"children":131},3,[132],{"text":133,"type":29},"为什么每个 context 配了 proxy 还是走同一个出口？",{"type":25,"children":135},[136],{"text":137,"type":29},"先检查浏览器启动参数和 context 配置，再核对代理侧的分配记录。如果多个 context 使用同一代理接入或共享同一会话，实际出口可能相同；Cookie 和 storage 隔离本身并不保证出口不同。",{"type":31,"level":130,"children":139},[140],{"text":141,"type":29},"context.request 会走 context 的代理吗？",{"type":25,"children":143},[144],{"text":145,"type":29},"会，但容易被漏测，需要单独断言。","",[148,282,406],{"source":6,"id":149,"documentId":150,"translationKey":150,"locale":9,"slug":151,"path":152,"_path":152,"title":153,"description":154,"excerpt":154,"image":155,"publishedAt":156,"updatedAt":157,"author":17,"tags":158,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":159,"bodyHtml":146},3987,"ehyrf4ygtsshhwkd0o87xat6","tls-certificate-compatibility-proxy-testing","\u002Fblog\u002Ftls-certificate-compatibility-proxy-testing","应用发布前，如何检查代理环境下的 TLS 证书兼容性","先分清直连、隧道转发和代理中间人三类链路，再逐级核对证书链、信任库位置与 SAN；本机正常而容器失败的，多为 CA 没进镜像。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_TLS_cover_ad09ec52a5.webp","2026-11-06T07:43:24.293Z","2026-11-06T07:43:24.278Z",[],[160,171,175,179,187,191,195,199,203,207,211,215,219,223,227,231,235,239,250,254,263,266,270,274,278],{"type":25,"children":161},[162,164,169],{"text":163,"type":29},"代理环境下的 TLS 问题，先要分清证书由谁签发：直连看目标证书，隧道转发看目标证书，代理中间人看代理自签 CA。",{"text":165,"type":29,"marks":166},"代理IP",[167],{"type":168},"bold",{"text":170,"type":29}," 本身不改变 TLS 握手，但企业代理或抓包工具会替换证书链，发布前必须把这条链路完整验一遍。",{"type":31,"level":21,"children":172},[173],{"text":174,"type":29},"先确认代理是否会替换目标站的 TLS 证书",{"type":25,"children":176},[177],{"text":178,"type":29},"直连或 HTTP CONNECT 隧道通常保留目标站点的证书，应检查证书链、有效期和 SAN（证书允许使用的域名）。如果企业代理会解密并重新签发 TLS 流量，客户端看到的是企业代理签发的站点证书；此时要确认该证书能连到受信任的企业 CA。先判断代理是否替换了证书，再确定检查对象。",{"type":25,"children":180},[181,183,186],{"text":182,"type":29},"核对代理协议与客户端接入方式时，可参考",{"text":56,"type":29,"marks":184},[185],{"href":59,"type":60},{"text":81,"type":29},{"type":31,"level":21,"children":188},[189],{"text":190,"type":29},"使用 OpenSSL 查看代理连接返回的证书链",{"type":25,"children":192},[193],{"text":194,"type":29},"```bash\nopenssl s_client -proxy proxy.example.com:8080 \\\n-connect target.example.com:443 \\\n-servername target.example.com -showcerts \\\n-verify_hostname target.example.com -verify_return_error \\\n-CAfile \u002Fpath\u002Fto\u002Fca-bundle.pem\n```",{"type":25,"children":196},[197],{"text":198,"type":29},"查看输出中的 `issuer`、有效期和 SAN，并重点看两件事：证书链能否构建到受信锚、主机名是否匹配。服务端通常只提供叶证书和中间证书，不要求它回传根证书；根证书应该来自你信任的本地信任库（`-CAfile` 或 `-CApath`），而不是从连接里“捡”到一个就信。验收看退出码和验证输出（`Verify return code`），不要只看“连接成功”；最后仍在真实应用运行时或容器里做一次正负对照。",{"type":25,"children":200},[201],{"text":202,"type":29},"`-proxy` 参数需要 OpenSSL 1.1.0 或更高版本，且默认只支持 HTTP CONNECT；SOCKS5 场景改用支持该协议的工具（如 curl 的 `--proxy socks5h:\u002F\u002F`）或对应的客户端验证。运行 `openssl version` 可先检查本机版本。",{"type":31,"level":21,"children":204},[205],{"text":206,"type":29},"检查系统、运行时和容器使用的证书信任库",{"type":25,"children":208},[209],{"text":210,"type":29},"信任库位置取决于应用使用的运行时：Linux 系统证书通常位于 `\u002Fetc\u002Fssl\u002Fcerts`，Java 使用自己的 `cacerts`，Node 可通过 `NODE_EXTRA_CA_CERTS` 指定额外 CA，Python 的 `requests` 常使用 `certifi`。容器还要单独检查镜像内的证书。本机正常而容器失败时，先比较两边实际加载的 CA。",{"type":31,"level":21,"children":212},[213],{"text":214,"type":29},"根据 TLS 错误定位证书链和主机名问题",{"type":25,"children":216},[217],{"text":218,"type":29},"出现 `unable to get local issuer certificate` 时，先检查证书链是否缺少中间 CA；出现 `self signed certificate` 时，确认它是否为预期的企业 CA 或测试证书。`hostname mismatch` 通常表示证书 SAN 不包含目标域名；`protocol version` 或 `cipher mismatch` 则需要核对客户端与服务端支持的 TLS 版本和加密套件。",{"type":31,"level":21,"children":220},[221],{"text":222,"type":29},"发布前检查证书有效期、信任链和轮换结果",{"type":31,"level":130,"children":224},[225],{"text":226,"type":29},"在预发布环境验证证书更新和轮换",{"type":25,"children":228},[229],{"text":230,"type":29},"发布前记录证书到期时间，证书轮换后复测一次；灰度环境的信任库要和生产一致，否则灰度通过、生产失败。",{"type":31,"level":130,"children":232},[233],{"text":234,"type":29},"记录证书签发者、有效期和信任库位置",{"type":25,"children":236},[237],{"text":238,"type":29},"发布前记录目标域名和端口、证书签发者与到期时间、信任库位置，以及代理是否会重新签发证书。\n检查清单四项：链能构建到受信锚、CA 在信任库、主机名匹配 SAN、证书未过期。任一项不满足就不能发布；发现证书替换影响生产流量时，立即回滚配置。",{"type":25,"children":240},[241,243,248],{"text":242,"type":29},"证书链异常仍无法归因时，可从",{"text":244,"type":29,"marks":245},"全球代理帮助中心",[246],{"href":247,"type":60},"https:\u002F\u002Fhelp.quanqiudaili.com\u002F",{"text":249,"type":29},"查找对应接入文档。",{"type":25,"children":251},[252],{"text":253,"type":29},"OpenSSL 参数以手册为准：",{"type":107,"format":108,"children":255},[256],{"type":111,"children":257},[258],{"text":259,"type":29,"marks":260},"`openssl s_client` 手册（`-proxy`、`-verify_hostname`、`-verify_return_error`）",[261],{"href":262,"type":60},"https:\u002F\u002Fdocs.openssl.org\u002F3.5\u002Fman1\u002Fopenssl-s_client\u002F",{"type":31,"level":21,"children":264},[265],{"text":128,"type":29},{"type":31,"level":130,"children":267},[268],{"text":269,"type":29},"代理环境下证书报错一定是代理问题吗？",{"type":25,"children":271},[272],{"text":273,"type":29},"不一定。先看证书链和信任库，主机名不符或 CA 缺失更常见。",{"type":31,"level":130,"children":275},[276],{"text":277,"type":29},"为什么本机正常、容器里失败？",{"type":25,"children":279},[280],{"text":281,"type":29},"容器镜像没有导入同一个 CA，信任库不一致。",{"source":6,"id":283,"documentId":284,"translationKey":284,"locale":9,"slug":285,"path":286,"_path":286,"title":287,"description":288,"excerpt":288,"image":289,"publishedAt":290,"updatedAt":291,"author":17,"tags":292,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":293,"bodyHtml":146},3986,"g5ee4zs0pa8hyuc307nk5nwf","proxy-error-handling-mock-testing","\u002Fblog\u002Fproxy-error-handling-mock-testing","如何在不依赖真实代理服务时测试应用的代理错误处理","用可注入的代理桩和故障矩阵覆盖超时、拒绝、认证失败，在受控环境验证重试、退避与降级；不接真实代理，也不记录任何真实凭据。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_cover_5ce09c130c.webp","2026-11-06T07:42:03.722Z","2026-11-06T07:42:03.708Z",[],[294,303,307,311,315,319,323,327,331,335,339,350,354,358,367,371,387,390,394,398,402],{"type":25,"children":295},[296,298,301],{"text":297,"type":29},"代理错误处理最适合用本地 mock 验证。407、超时、连接重置、TLS 失败、慢响应、半关闭六类故障都能稳定复现，测试不再依赖服务商侧的偶发故障。",{"text":165,"type":29,"marks":299},[300],{"type":168},{"text":302,"type":29}," 的真实链路不可控，mock 的价值就是把“偶发”变成“可重复”，但结论只对应用逻辑成立。",{"type":31,"level":21,"children":304},[305],{"text":306,"type":29},"代理错误测试需要覆盖的六种故障",{"type":25,"children":308},[309],{"text":310,"type":29},"可先覆盖六种常见故障：让 mock 返回带 `Proxy-Authenticate` 的 407，检查客户端是否重试或报告认证错误；延迟超过客户端超时时，检查超时处理；连接重置时，检查中断后的动作；TLS 负例一次只改一个变量（未受信 CA、可信 CA 但 SAN 不对、证书过期），并保持校验开启，看客户端实际报什么错；分段延迟发送时，记录首字节和完整响应耗时；测“响应不完整”时，用声明了 `Content-Length` 却提前 EOF、或缺少 chunked 终止块的响应做负例，完整的 close-delimited 响应只作正常对照——单纯关闭写方向并不等于响应不完整。之后再按应用实际支持的错误分支补充测试。\n矩阵按应用实际支持的错误分支建立，上表是起步集合，可以补代理 DNS 失败、认证挑战格式异常和重试上限三类用例。矩阵未覆盖的分支，不会被这一组测试验证。",{"type":31,"level":21,"children":312},[313],{"text":314,"type":29},"如何搭建可重复触发故障的 mock 代理",{"type":25,"children":316},[317],{"text":318,"type":29},"mitmproxy、toxiproxy 或自建 TCP 服务都可以用于 mock，但注入点要和故障所在层对应。407 需要能处理 CONNECT 并返回 `Proxy-Authenticate` 的受控代理；TCP 延迟和连接重置可在传输层注入；SOCKS 认证错误则要按 SOCKS 协议构造，不能用 HTTP 的 407 代替。记录工具、客户端版本、触发步骤和清理方法，其他人才能复现。",{"type":31,"level":21,"children":320},[321],{"text":322,"type":29},"除了错误码，还要验证应用如何处理故障",{"type":25,"children":324},[325],{"text":326,"type":29},"收到 407 只说明应用识别到了代理认证错误。还要检查重试次数和间隔、熔断或降级是否符合预期，以及日志是否隐藏凭据。407 和 401 要分开判断：前者由代理要求认证，后者由目标服务要求认证。",{"type":25,"children":328},[329],{"text":330,"type":29},"407 和 401 必须分开断言：`Proxy-Authenticate` 是代理要求认证，`WWW-Authenticate` 是目标服务要求认证。混在一起，会把代理认证失败记成应用登录失败。",{"type":31,"level":21,"children":332},[333],{"text":334,"type":29},"避免 mock 测试漏掉真实故障分支",{"type":25,"children":336},[337],{"text":338,"type":29},"要触发客户端超时，mock 延迟必须超过客户端配置的超时值；要测试请求中断后的处理，应在连接建立后再重置。mock 结果只说明应用对模拟故障的处理方式，不能直接推断真实代理链路表现。",{"type":25,"children":340},[341,343,348],{"text":342,"type":29},"mock 通过后若要转入真实环境验证，可按",{"text":344,"type":29,"marks":345},"全球代理使用流程",[346],{"href":347,"type":60},"https:\u002F\u002Fhelp.quanqiudaili.com\u002Fdocs\u002FProcess",{"text":349,"type":29},"依次核对产品、参数和接入步骤。",{"type":31,"level":21,"children":351},[352],{"text":353,"type":29},"记录 mock 故障参数和应用处理结果",{"type":25,"children":355},[356],{"text":357,"type":29},"每个用例记录故障类型、注入参数、期望动作、实际结果，以及日志是否隐藏凭据。\n用例边界是应用逻辑，不覆盖真实服务商的 SLA 和地区行为。出现真实凭据进入 mock 日志、或测试影响到共享环境时，立即停止并清理。",{"type":25,"children":359},[360,362,365],{"text":361,"type":29},"开始真实环境验证前，可用",{"text":56,"type":29,"marks":363},[364],{"href":59,"type":60},{"text":366,"type":29},"核对协议、地址和认证参数。",{"type":25,"children":368},[369],{"text":370,"type":29},"故障注入工具与协议定义：",{"type":107,"format":108,"children":372},[373,380],{"type":111,"children":374},[375],{"text":376,"type":29,"marks":377},"toxiproxy（TCP 延迟、重置、带宽限制）",[378],{"href":379,"type":60},"https:\u002F\u002Fgithub.com\u002FShopify\u002Ftoxiproxy",{"type":111,"children":381},[382],{"text":383,"type":29,"marks":384},"RFC 9110《HTTP Semantics》（407 与 `Proxy-Authenticate`）",[385],{"href":386,"type":60},"https:\u002F\u002Fwww.rfc-editor.org\u002Frfc\u002Frfc9110.html",{"type":31,"level":21,"children":388},[389],{"text":128,"type":29},{"type":31,"level":130,"children":391},[392],{"text":393,"type":29},"mock 代理能替代真实代理测试吗？",{"type":25,"children":395},[396],{"text":397,"type":29},"不能。它验证应用逻辑，不验证真实链路的稳定性和地区行为。",{"type":31,"level":130,"children":399},[400],{"text":401,"type":29},"407 和 401 有什么区别？",{"type":25,"children":403},[404],{"text":405,"type":29},"407 是代理要求认证，401 是目标服务要求认证。",{"source":6,"id":407,"documentId":408,"translationKey":408,"locale":9,"slug":409,"path":410,"_path":410,"title":411,"description":412,"excerpt":412,"image":413,"publishedAt":414,"updatedAt":415,"author":17,"tags":416,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":417,"bodyHtml":146},3985,"bpamp8c4gnpz3p9kslssa97y","ci-runner-proxy-build-impact-testing","\u002Fblog\u002Fci-runner-proxy-build-impact-testing","如何在 CI Runner 中验证代理不会影响构建步骤","在 Runner 内固定镜像、构建输入与测试时段，用出口 IP、DNS、TLS 和依赖清单作为证据，区分代理影响与构建环境波动；产物哈希只在构建可复现时才作对照。","https:\u002F\u002Fcms.quanqiudaili.com\u002Fuploads\u002Flarge_CI_Runner_cover_023a325e36.webp","2026-11-06T07:41:47.066Z","2026-11-06T07:41:47.049Z",[],[418,422,426,430,434,438,447,451,455,459,463,467,471,475,484,488,504,507,511,515,519],{"type":25,"children":419},[420],{"text":421,"type":29},"判断代理是否影响 CI，不能只看最终构建是否成功。应按实际流程检查代码拉取、依赖安装、单元测试、构建和产物上传。代理配置可能来自环境变量、包管理器或容器后台服务；配置不一致时，本地命令与流水线可能表现不同。",{"type":31,"level":21,"children":423},[424],{"text":425,"type":29},"如何判断代理没有影响 CI 构建",{"type":25,"children":427},[428],{"text":429,"type":29},"按顺序检查代码是否拉取到相同 commit、依赖校验是否通过、单元测试结果是否与基线一致。构建可复现时比较产物哈希；否则比较关键文件和依赖清单。上传环节要确认远端文件实际存在，并在支持时核对摘要或完整性校验。各阶段记录退出码和耗时；确需记录出口 IP 时，应使用受控日志并隐藏代理凭据。",{"type":31,"level":21,"children":431},[432],{"text":433,"type":29},"检查 CI 环境变量、包管理器和容器的代理配置",{"type":25,"children":435},[436],{"text":437,"type":29},"先检查构建进程是否读取 `HTTP_PROXY` 和 `HTTPS_PROXY`，再确认 `NO_PROXY` 中哪些地址会绕过代理。npm、pip 和 Maven 也可能分别读取 `.npmrc`、`pip.conf` 或 `settings.xml`；容器拉取镜像则由容器后台服务的配置决定。不要只凭宿主机环境变量判断代理已经生效。",{"type":25,"children":439},[440,442,445],{"text":441,"type":29},"配置入口和协议参数可对照",{"text":56,"type":29,"marks":443},[444],{"href":59,"type":60},{"text":446,"type":29},"逐项核实。",{"type":31,"level":21,"children":448},[449],{"text":450,"type":29},"对比启用代理和不启用代理时的构建结果",{"type":25,"children":452},[453],{"text":454,"type":29},"在相同 commit、Runner 镜像、依赖来源和资源规格下，分别运行启用代理与不启用代理的任务。冷缓存和暖缓存应分开比较，并使用隔离的缓存目录；每组可重复运行并保留原始结果。除代理开关外，其他条件尽量保持一致，便于判断差异来源。",{"type":31,"level":21,"children":456},[457],{"text":458,"type":29},"区分正常的耗时变化与构建产物差异",{"type":25,"children":460},[461],{"text":462,"type":29},"代理会让下载耗时发生变化，但不能预设“代理一定更慢”。判断是否影响构建，用各阶段耗时分布和事前设定的预算，不用单次耗时下结论；产物对照的前提是构建可复现、非确定性元数据已规范化；做不到时改比依赖清单、测试结果、关键文件和构建日志，不能仅凭整体哈希变化判定代理造成回归。",{"type":25,"children":464},[465],{"text":466,"type":29},"两类误判最常见：`NO_PROXY` 漏配，内部制品库被送进代理，出现 407 或超时；缓存掩盖失败，上一次的缓存让本次跳过了真实下载，清缓存复跑才暴露问题。",{"type":31,"level":21,"children":468},[469],{"text":470,"type":29},"记录 CI 代理测试结果并设定验收条件",{"type":25,"children":472},[473],{"text":474,"type":29},"记录 Runner 镜像、commit、实际生效的代理配置和各阶段结果，并注明缓存命中情况。被缓存跳过的步骤应标为未覆盖，必要时单独补跑。所有相关阶段通过后，再依据构建是否可复现判断产物差异。若影响其他流水线、日志出现凭据或耗时超过预算，应停止并回滚配置。",{"type":25,"children":476},[477,479,482],{"text":478,"type":29},"CI 中出现连接失败时，可从",{"text":244,"type":29,"marks":480},[481],{"href":247,"type":60},{"text":483,"type":29},"查找官方接入说明。",{"type":25,"children":485},[486],{"text":487,"type":29},"相关机制说明：",{"type":107,"format":108,"children":489},[490,497],{"type":111,"children":491},[492],{"text":493,"type":29,"marks":494},"`npm ci` 的安装与校验行为",[495],{"href":496,"type":60},"https:\u002F\u002Fdocs.npmjs.com\u002Fcli\u002Fv10\u002Fcommands\u002Fnpm-ci",{"type":111,"children":498},[499],{"text":500,"type":29,"marks":501},"GitHub Actions 依赖缓存",[502],{"href":503,"type":60},"https:\u002F\u002Fdocs.github.com\u002Factions\u002Fusing-workflows\u002Fcaching-dependencies-to-speed-up-workflows",{"type":31,"level":21,"children":505},[506],{"text":128,"type":29},{"type":31,"level":130,"children":508},[509],{"text":510,"type":29},"CI 里设了 HTTP_PROXY 为什么还是不走代理？",{"type":25,"children":512},[513],{"text":514,"type":29},"可能被 NO_PROXY 覆盖，或包管理器有独立配置，以实际生效的配置为准。",{"type":31,"level":130,"children":516},[517],{"text":518,"type":29},"构建成功就说明代理没影响吗？",{"type":25,"children":520},[521],{"text":522,"type":29},"不一定。缓存可能掩盖下载失败，需要清缓存复跑并比对产物。",[524],{"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":525,"featured":19,"sortOrder":20,"readingTime":21,"bodyType":22,"bodyBlocks":526,"bodyHtml":146},[],[527,530,533,536,539,542,545,552,555,558,565,568,571,574,577,580,583,595,598,601,604,607],{"type":25,"children":528},[529],{"text":28,"type":29},{"type":31,"level":21,"children":531},[532],{"text":34,"type":29},{"type":25,"children":534},[535],{"text":38,"type":29},{"type":25,"children":537},[538],{"text":42,"type":29},{"type":25,"children":540},[541],{"text":46,"type":29},{"type":31,"level":21,"children":543},[544],{"text":50,"type":29},{"type":25,"children":546},[547,548,551],{"text":54,"type":29},{"text":56,"type":29,"marks":549},[550],{"href":59,"type":60},{"text":62,"type":29},{"type":25,"children":553},[554],{"text":66,"type":29},{"type":31,"level":21,"children":556},[557],{"text":70,"type":29},{"type":25,"children":559},[560,561,564],{"text":74,"type":29},{"text":76,"type":29,"marks":562},[563],{"href":79,"type":60},{"text":81,"type":29},{"type":31,"level":21,"children":566},[567],{"text":85,"type":29},{"type":25,"children":569},[570],{"text":89,"type":29},{"type":31,"level":21,"children":572},[573],{"text":93,"type":29},{"type":25,"children":575},[576],{"text":97,"type":29},{"type":25,"children":578},[579],{"text":101,"type":29},{"type":25,"children":581},[582],{"text":105,"type":29},{"type":107,"format":108,"children":584},[585,590],{"type":111,"children":586},[587],{"text":114,"type":29,"marks":588},[589],{"href":117,"type":60},{"type":111,"children":591},[592],{"text":121,"type":29,"marks":593},[594],{"href":124,"type":60},{"type":31,"level":21,"children":596},[597],{"text":128,"type":29},{"type":31,"level":130,"children":599},[600],{"text":133,"type":29},{"type":25,"children":602},[603],{"text":137,"type":29},{"type":31,"level":130,"children":605},[606],{"text":141,"type":29},{"type":25,"children":608},[609],{"text":145,"type":29},null,[],[]]