[{"data":1,"prerenderedAt":216},["ShallowReactive",2],{"blog:slow-at-night-peak-hours":3},{"zh-CN":4,"en":116},{"blocks":5,"faq":97},[6,9,11,13,15,19,21,26,28,31,33,37,39,41,44,46,48,51,53,64,66,69,71,80,83,85,90,93,95],{"type":7,"html":8},"p","同一个节点，同一台设备，白天用得挺顺，一到晚上就开始卡——这种规律性的变慢，多数时候不是 VPN 坏了，是晚高峰期间某一段线路挤满了人。搞清楚堵在哪一段，比把节点换个十几遍更有用，也比干等到半夜网速自己恢复更省时间。",{"type":7,"html":10},"排查原则不变：\u003Cstrong>一次只改一个变量\u003C\u002Fstrong>。同时换节点、换设备、又切网络，问题好转了也不知道是哪一步起的作用。",{"type":7,"html":12},"「VPN 坏了」这个判断听起来简单，但站不住脚——同一个客户端、同一个账号，白天用得好好的，坏的东西从来不会只在特定几个小时坏，会一直坏。真正会按时间段变化的，只有网络这条路径上的拥堵程度。",{"type":7,"html":14},"这类规律性的变慢有个特点：症状每天差不多在同一个时间段出现，白天基本正常。抓住「固定时间段」这个特征，就已经排除了设备坏了、账号出问题这类偶发性原因，方向直接缩小到跟拥堵有关的几种可能。",{"type":16,"id":17,"html":18},"h2","bp-triage","先做一次判断：不开 VPN 单独测一次速度",{"type":7,"html":20},"在同一个网络、同一个变慢的时间段，先关掉 VPN 单独测一次速度，再开着 VPN 测一次，把两个数字都记下来。",{"type":22,"items":23},"list",[24,25],"不开 VPN 也慢：问题出在你这边的网络本身，往下看「最后一公里」那一节，你这一头的部分。","不开 VPN 很快，开了 VPN 才慢：问题出在 VPN 这条路径的某一段，继续往下查节点或者更大范围的拥堵。",{"type":7,"html":27},"这一步能直接把「自己家网络的问题」和「VPN 路径上的问题」分成两条线，避免两边一起怀疑。很多人卡了第一反应就是换节点，换了一圈发现还是慢，才想起来测一下不开 VPN 是不是也一样——这个顺序反了，先测这一步能省掉后面一大半的折腾，两次测速加起来也就一两分钟。",{"type":16,"id":29,"html":30},"bp-lastmile","最常见：晚高峰堵的是「最后一公里」，不是 VPN 坏了",{"type":7,"html":32},"不管用不用 VPN，「最后一公里」指的都是你的设备连到你的网络服务商之间那一小段。这一段在晚高峰期间几乎必然更挤，原因分两头看。",{"type":34,"id":35,"html":36},"h3","sub-1","你这一头的最后一公里",{"type":7,"html":38},"同一个小区、同一栋楼里的用户，晚上下班回家以后几乎同一时间开始看视频、打游戏、下载更新，而这些用户往往共享同一段本地线路的上行带宽。用的人一多，分到每一户的带宽自然变少，这跟有没有开 VPN 完全没关系——不开 VPN 一样会慢，只是平时不太会专门去测，直到某天开着 VPN 觉得卡，才第一次注意到这个时间段的网络本来就不太好。",{"type":7,"html":40},"这类拥堵还有个规律：越晚高峰时间越集中的居住区，比如大量年轻租户聚集的小区，晚间的拥堵往往比商业区附近更明显，因为后者晚上在线设备数量本身就少。",{"type":34,"id":42,"html":43},"sub-2","出口那一头的最后一公里",{"type":7,"html":45},"VPN 节点自己连到更大范围网络的那一段，也有类似的「最后一公里」，一样会在使用人数最多的时候变挤。关键是这个「晚高峰」按的是节点所在地区的时钟，不是你自己的时钟——如果节点在跟你时区差很多的地方，它的高峰期可能出现在你这边完全不相关的时间点，也可能刚好跟你自己的晚上重叠，纯粹看地区。",{"type":7,"html":47},"举个例子：如果一个节点所在的地区和你自己所在的地方时差不大，两边的晚高峰基本重叠，你的晚上恰好也是它的晚上，拥堵会叠加在一起；换到一个时差有七八个小时以上的地区，你的晚上落在对方的白天甚至凌晨，反而更容易避开真正拥挤的那一段窗口。",{"type":16,"id":49,"html":50},"bp-other-congestion","其次：赶上更大范围的规律性拥堵",{"type":7,"html":52},"除了两头的最后一公里，中间连接不同网络的骨干线路，也存在固定时间段拥堵的情况——这类拥堵影响的不是一个节点，而是经过那条骨干线路的一大片用户，表现上跟单个节点出问题很像，但换节点解决不了，因为换到的下一个节点很可能走的还是同一条骨干。",{"type":54,"items":55},"steps",[56,60],{"title":57,"body":58},"怎么确认",[59],"换两三个不同地区的节点，如果晚高峰这几个点全都变慢，白天测同样几个点都正常，基本可以判断是这类更大范围的拥堵，不是某一个节点的问题。",{"title":61,"body":62},"怎么处理",[63],"这类拥堵没有稳定的绕开办法，能做的是错开真正的高峰时间段——提前半小时或者延后半小时，很多时候就能避开最挤的那一小段窗口。同一时间试试跨时区更远的节点，有时候比在相近地区里来回换更有效，原理跟上一节的时钟错位是一样的。",{"type":65},"downloads",{"type":16,"id":67,"html":68},"bp-node-crowd","少见：不是时间问题，是那个节点长期人就多",{"type":7,"html":70},"有些节点的问题不是晚高峰才卡，是本身长期承载的用户就偏多，晚高峰只是让本来就紧张的情况更明显。这种情况换个时间段测也好不了太多，因为瓶颈是长期性的，不是某个时间点才出现的，白天看着凑合，只是因为白天的基数本来就比晚上低一些。",{"type":54,"items":72},[73,76],{"title":57,"body":74},[75],"挑一个平常日子的非高峰时段，专门测一下这个节点，如果连白天效果都比其他节点明显差，就是这个节点本身承载偏多，不是晚高峰的锅，换个节点对比一下差距有多大最直观。",{"title":61,"body":77},[78,79],"把常用节点换成另一个地区，作为日常默认使用，而不是只在晚上卡的时候临时换一下——长期性的问题需要长期换一个更稳的默认选项，而不是每天临时补救。","判断一个节点是不是长期超载，观察几天比测一次更准。同一个地区如果有好几个可选节点，轮着用几天记一下感觉，比只凭一次测速下结论靠谱。",{"type":16,"id":81,"html":82},"bp-other-factors","排除设备和家里其他因素的干扰",{"type":7,"html":84},"拥堵之外，晚上这几个小时还容易叠加几个跟拥堵无关的因素：",{"type":22,"items":86},[87,88,89],"家里其他人也在晚高峰这几个小时上网，同一个路由器要同时应付好几台设备，家用路由器本身撑不住的时候，表现跟线路拥堵很像，测速的时候最好问一下家里其他人是不是也在同时下载或者看视频。","路由器用久了，晚上负载最高的这几个小时最容易暴露性能不够，重启一次能不能缓解，是判断路由器是不是瓶颈的简单办法——能缓解说明路由器本身撑不住高峰负载，不能缓解才往网络这一层继续查。","视频或者游戏本身在晚间高峰也会遇到服务器那一端的压力，跟网络这一层是两个独立的环节，\u003Ca href=\"\u002Fblog\u002Fvideo-buffering-with-vpn\u002F\">视频一直缓冲\u003C\u002Fa>那篇专门拆过这条链，两篇结合起来看能覆盖大部分晚上变慢的情况。",{"type":16,"id":91,"html":92},"bp-final","总结",{"type":7,"html":94},"一到晚上就变慢，先用不开 VPN 单独测速这一步判断问题在你这边还是在 VPN 路径上，再看是两头的最后一公里、更大范围的骨干拥堵，还是节点本身长期人多，别漏了路由器和家里其他设备这类容易被忽略的干扰。真正需要长期解决的，是给自己留一个不那么拥挤的默认节点，而不是每天晚上重新试一遍。",{"type":7,"html":96},"如果几个节点来回换还是找不到规律，\u003Ca href=\"\u002Fblog\u002Fvpn-slow-speed-fixes\u002F\">VPN 速度慢：先查这六个地方\u003C\u002Fa>覆盖的范围更全；测速数字和实际体验对不上的情况，另有\u003Ca href=\"\u002Fblog\u002Fwhy-speed-test-looks-wrong\u002F\">为什么测速数字好看实际却很卡\u003C\u002Fa>可以看。\u003Ca href=\"\u002Ffeatures\u002F\">快橙\u003C\u002Fa>换节点只要一步，遇到某条线路晚高峰变慢，几秒钟就能换到另一条上试，注册后每天有免费试用时长可以先测起来。",[98,101,104,107,110,113],{"q":99,"a":100},"白天用得好好的，一到晚上就卡，是不是 VPN 本身不行？","大部分时候是晚高峰期间某一段线路拥堵，不是 VPN 坏了。不开 VPN 单独测速，同样在这个时间段变慢，就能证明跟 VPN 没关系。",{"q":102,"a":103},"换了好几个节点，晚上都一样卡，为什么？","说明拥堵可能不在某一个具体节点上，而是更大范围的晚高峰。这种情况下换节点效果有限，换个不同的时间段测一次更能看出差别。",{"q":105,"a":106},"是不是节点选的地区不对？","节点所在地区的晚高峰，未必和你自己这边的晚上是同一段时间。可以试着换一个跟自己时区差异比较大的节点，看拥堵是不是明显不一样。",{"q":108,"a":109},"晚高峰一般是几点到几点？","因地区和网络而异，没有统一的数字，大致是当地下班到睡前那几个小时。自己在不同时间点各测一次，比套用一个通用时间段准确。",{"q":111,"a":112},"换成有线网络能避开晚高峰吗？","能避开自己这头的一部分变量。移动网络的基站在人多的晚间本身也会拥堵，有线通常更稳，两种拥堵叠加的时候，先排除自己这头会更快定位问题。",{"q":114,"a":115},"快橙的节点晚上会不会也很挤？","只要有人用就有拥堵的可能，快橙能做的是让换节点只要一步，遇到某条线路晚高峰变慢，几秒钟就能换到另一条上试。",{"blocks":117,"faq":197},[118,120,122,124,126,128,130,134,136,138,140,143,145,147,150,152,154,156,158,168,169,171,173,182,184,186,191,193,195],{"type":7,"html":119},"Same node, same device, smooth all day, and slow the moment evening hits — this kind of clockwork slowdown is rarely the VPN breaking. It's one segment of the path filling up with people during peak hours. Finding which segment is congested beats switching nodes a dozen times, and it beats waiting until midnight for speed to recover on its own.",{"type":7,"html":121},"The same rule as always: \u003Cstrong>change one variable at a time\u003C\u002Fstrong>. Switch the node, switch the device and switch networks all at once, and even once things improve you won't know which step actually did it.",{"type":7,"html":123},""The VPN is broken" sounds like a simple explanation, but it doesn't hold up — the same client, the same account, working fine all day. Something broken doesn't break for only a few specific hours and work the rest of the time; it just stays broken. The one thing that genuinely changes by time of day is how congested this particular path happens to be.",{"type":7,"html":125},"This kind of clockwork slowdown has a tell: the symptom shows up at roughly the same time every day, and the daytime is fine. That one detail — a fixed time window — already rules out a broken device or a broken account, both of which would misbehave at random, and narrows things down to a short list of congestion-related causes.",{"type":16,"id":17,"html":127},"Start with one test: run a speed test with the VPN off",{"type":7,"html":129},"On the same network, at the same hour things usually slow down, run a speed test with the VPN off, then run one with it on, and write down both numbers.",{"type":22,"items":131},[132,133],"Slow even without the VPN: the problem is on your own network — keep reading for the "last mile," specifically the part on your end.","Fast without the VPN, slow only with it on: the problem is somewhere along the VPN's own path — keep reading to check the node or wider congestion.",{"type":7,"html":135},"This one step cleanly separates "a problem with my own network" from "a problem somewhere on the VPN's path," instead of suspecting both at once. The usual first move is to switch nodes, discover it's still slow, and only then think to test with the VPN off — that's backwards. Testing this first saves most of what comes after, and the two tests together take barely a minute or two.",{"type":16,"id":29,"html":137},"Most common: peak hours congest the "last mile," not the VPN itself",{"type":7,"html":139},"Whether a VPN is involved or not, the "last mile" is the short segment between your device and your own internet provider. It gets busier during evening peak hours almost everywhere, for two separate reasons.",{"type":34,"id":141,"html":142},"the-last-mile-on-your-end","The last mile on your end",{"type":7,"html":144},"Everyone in the same building or neighbourhood gets home from work at roughly the same time and starts streaming, gaming or downloading updates almost simultaneously, and these users often share the upstream capacity of the same local line. More people using it means less bandwidth for each household, and that has nothing to do with whether a VPN is running — it would be just as slow without one, only nobody usually bothers to check until the day it feels slow with the VPN on and they notice for the first time that this hour was never great to begin with.",{"type":7,"html":146},"There's a pattern to this too: neighbourhoods with more concentrated evening usage, like ones with a lot of young renters, tend to show worse evening congestion than areas near business districts, simply because the second kind has fewer devices online at night.",{"type":34,"id":148,"html":149},"the-last-mile-on-the-exit-end","The last mile on the exit end",{"type":7,"html":151},"The VPN node has its own "last mile" connecting it to the wider internet, and it gets just as congested when usage peaks. The key detail is that this "peak hour" runs on the clock of wherever the node sits, not yours — if the node is in a region with a very different time zone, its busiest hours might land at a time that means nothing to you, or they might land squarely on your own evening. It comes down entirely to the region.",{"type":7,"html":153},"Here's the logic: if a node sits in a region close to your own time zone, the two evening peaks mostly line up, so your evening is also its evening and the congestion stacks. Switch to a region seven or eight hours apart or more, and your evening lands somewhere in that region's daytime or even the middle of its night — which makes it easier to land outside the truly crowded window.",{"type":16,"id":49,"html":155},"Next most common: a wider, regularly recurring congestion",{"type":7,"html":157},"Beyond the last mile on either end, the backbone links connecting different networks in the middle can also get congested at fixed times of day. This kind of congestion doesn't affect one node — it affects everyone passing through that backbone link, which looks a lot like a single node having a problem, except switching nodes doesn't fix it, because the next node very likely routes through the same backbone anyway.",{"type":54,"items":159},[160,164],{"title":161,"body":162},"How to confirm it",[163],"Try two or three nodes in different regions. If all of them slow down at the same peak hours and all of them test fine during the day, that points at this wider kind of congestion rather than any one node.",{"title":165,"body":166},"How to fix it",[167],"There's no reliable way around this kind of congestion — the practical move is shifting off the real peak window, half an hour earlier or later is often enough to miss the worst of it. Trying a node much further away in time zone at the same moment is also worth doing, and it can work better than cycling through nearby regions, for the same time-zone-mismatch reason covered above.",{"type":65},{"type":16,"id":67,"html":170},"Rarer: not a time-of-day issue, the node is just consistently crowded",{"type":7,"html":172},"Some nodes aren't slow only at peak hours — they're carrying more users than they comfortably should on an ongoing basis, and the evening peak just makes an already-tight situation more obvious. Testing at a different time of day won't help much here, because the bottleneck is structural, not tied to a particular hour; daytime looks tolerable purely because the daytime baseline is lower than the evening one.",{"type":54,"items":174},[175,178],{"title":161,"body":176},[177],"Pick an off-peak moment on an ordinary day and test that specific node on its own. If even its daytime performance is noticeably worse than other nodes, the node itself is carrying too much load — the evening slowdown was never really about the evening. Testing a different node side by side makes the gap obvious.",{"title":165,"body":179},[180,181],"Switch your everyday default to a node in a different region, rather than only swapping temporarily when it feels slow at night — a structural problem needs a steadier default, not a nightly workaround.","Judging whether a node is chronically overloaded takes a few days of observation, not one test. If a region offers several nodes, rotating through them for a few days and keeping notes beats drawing a conclusion from a single speed test.",{"type":16,"id":81,"html":183},"Rule out the device and everything else at home",{"type":7,"html":185},"Beyond congestion, evening hours tend to stack a few unrelated factors on top:",{"type":22,"items":187},[188,189,190],"Other people at home are online during the same peak hours, and one router has to serve several devices at once — when a home router simply can't keep up, it looks a lot like line congestion. Worth asking whether anyone else is downloading or streaming at the same moment you're testing.","An older router is most likely to show its age during the busiest hours of the night. Whether a reboot helps is a simple way to tell — if it does, the router itself can't handle peak load; if it doesn't, keep looking at the network layer instead.","Video and games have their own server-side load during evening peaks too, a separate question from this network layer entirely — \u003Ca href=\"\u002Fen\u002Fblog\u002Fvideo-buffering-with-vpn\u002F\">video that keeps buffering\u003C\u002Fa> breaks that chain down on its own, and the two together cover most evening slowdowns.",{"type":16,"id":91,"html":192},"In summary",{"type":7,"html":194},"For a slowdown that hits every evening, start with a VPN-off speed test to tell whether the problem is on your end or somewhere on the VPN's path, then check the last mile on both ends, wider backbone congestion, or a node that's chronically overloaded — without overlooking the router and other devices at home, which are easy to miss. What actually needs fixing long-term is keeping a less congested node as your default, rather than re-testing from scratch every night.",{"type":7,"html":196},"If cycling through nodes still doesn't reveal a pattern, \u003Ca href=\"\u002Fen\u002Fblog\u002Fvpn-slow-speed-fixes\u002F\">VPN running slow: check these six things first\u003C\u002Fa> covers more ground; when a good speed test number doesn't match how it actually feels, \u003Ca href=\"\u002Fen\u002Fblog\u002Fwhy-speed-test-looks-wrong\u002F\">why a good speed test can still feel slow\u003C\u002Fa> is worth reading too. \u003Ca href=\"\u002Fen\u002Ffeatures\u002F\">FastOrange\u003C\u002Fa> switches nodes in one tap, so hitting a slow line at peak hours means trying another one within seconds — every registered account gets free daily trial time to start testing with.",[198,201,204,207,210,213],{"q":199,"a":200},"It works fine all day and slows down every evening. Is the VPN just bad?","Most of the time it's evening congestion on one segment of the path, not the VPN failing. If speed drops at that same hour with the VPN off too, that proves it has nothing to do with the VPN.",{"q":202,"a":203},"I've tried several nodes and they're all slow at night. Why?","That suggests the congestion isn't tied to one specific node — it's a wider peak-hour effect. Switching nodes won't help much here; testing at a different time of day shows the difference more clearly.",{"q":205,"a":206},"Could I just have picked the wrong region for the node?","A node's peak hour doesn't necessarily line up with your own evening. Try a node in a region with a much bigger time difference from yours and see whether the congestion looks noticeably different.",{"q":208,"a":209},"What hours count as peak, exactly?","There's no universal number — it depends on the region and the network, roughly the hours between when people get off work and go to bed locally. Testing at a few different times yourself beats assuming a fixed window.",{"q":211,"a":212},"Would a wired connection avoid the evening slowdown?","It rules out part of what's on your own end. Mobile towers get congested in busy evening hours too, and wired connections are usually steadier — when both kinds of congestion stack up, ruling out your own end first narrows things down faster.",{"q":214,"a":215},"Do FastOrange's own nodes get crowded at night too?","Any node with real usage can get congested. What FastOrange can do is make switching nodes take one tap, so hitting a slow line at peak hours is a matter of trying another one within seconds.",1793276292144]