Fluentide

A Chinese podcast · Same story, 4 levels

Microsoft's Email Went Down for a Day — episode cover art
2:09

微软的邮件,停了一天

Microsoft's Email Went Down for a Day

About this story

Outlook, Teams and other Microsoft services stopped for about a day in November 2026. Microsoft says a hidden bug was the cause. HSK 5+ Chinese listening practice.

This is an HSK 5+ Chinese listening episode that runs about 2 minutes. The full Mandarin script is shown with tap-for-pinyin and a line-by-line English translation, so you can listen and read at once — comprehensible input in the sense of Stephen Krashen's i+1 theory. It teaches 12 key vocabulary words such as 办法、结果、原因 and walks through 4 grammar patterns, each explained in English with examples. The same news story is retold at 4 difficulty levels — use the level selector above to find the version that is challenging but still understandable for you.

Read at your level

原文

Read the complete story in Chinese. Reveal pinyin and English only when you need them.

关掉再打开,是修电脑最老的办法,十一月 Microsoft 的工程师也试了,结果越修越糟。
十一月十二号中午前后,Outlook 突然连不上,Teams 和 Copilot 也跟着出问题,故障从美洲蔓延到全球,持续了将近一天。
Microsoft 很早就表示没有发现恶意活动,也就是说,它不认为这是一次攻击。
根据它公布的初步报告,真正的原因是三件事叠在了一起。
第一件,七月的一次更新,在 Microsoft 365 的网关服务器里埋下了一个漏洞,三个多月里一直没有发作。
第二件,出事前大约十五个小时,Teams 的流量被自动切换到了 Exchange Online 使用的网络上,这些服务器的压力一下子大了。
第三件,将近百分之十四的网关服务器,因为另一个问题已经下线。
Microsoft 说,少了这些机器,容量本来也够用,可是流量变多、机器变少,正好触发了那个漏洞。
接下来是一场连锁反应。
一台服务器失去反应,流量被转给别的服务器,别的服务器也跟着倒下。
用来防止故障跨地区扩散的保护措施,这次没有挡住。
不到两小时,这个漏洞在整个系统里发作,最严重的时候,超过八成的服务器没有反应。
这时候重启也没有用,服务器刚回来,积压的流量马上又把它压垮。
Teams 这些服务可以换到别的网络上,恢复得早一些,Exchange Online 没有这条退路,所以邮件停得最久。
Microsoft 还承认,平时的负载测试没有测出这个漏洞,而且机器越新、核心越多,问题越严重。
十三号,工程师给服务器装上修复补丁,邮件才陆续恢复。
Microsoft 向客户道歉,把邮件的重试时间延长到四十八小时,并且增加了容量和监控。
这让很多人想起二零二四年 CrowdStrike 的那次更新事故,一个小错误,同样让全世界的电脑停了下来。
完整的调查报告,Microsoft 说过几天公布。
English transcript reference

Turning it off and on again is the oldest way to fix a computer; in November Microsoft's engineers tried it too, and the more they fixed, the worse it got.

Around midday on 12 November, Outlook suddenly could not connect, Teams and Copilot ran into trouble as well, and the failure spread from the Americas to the whole world and lasted nearly a day.

Microsoft said early on that it had found no malicious activity; in other words, it does not think this was an attack.

According to the preliminary report it published, the real cause was three things piling up together.

The first: an update in July buried a flaw in the gateway servers of Microsoft 365, and for more than three months it never went off.

The second: about fifteen hours before the trouble, Teams traffic was switched automatically onto the network that Exchange Online uses, and the load on these servers suddenly grew.

The third: nearly fourteen percent of the gateway servers were already offline because of another problem.

Microsoft said that even without these machines there was enough capacity, but more traffic and fewer machines together triggered exactly that flaw.

What followed was a chain reaction.

One server stopped responding, its traffic was passed to other servers, and they went down as well.

The safeguards meant to stop a failure spreading across regions did not hold this time.

In less than two hours the flaw went off across the whole system, and at its worst more than eighty percent of the servers were not responding.

By then restarting was no use either: the moment a server came back, the backed-up traffic crushed it again.

Services like Teams could be moved to other networks and recovered earlier; Exchange Online had no such way out, so email was down the longest.

Microsoft also admitted that its usual load testing did not catch this flaw, and that the newer the machine and the more cores it had, the worse the problem was.

On the 13th, engineers installed a fix on the servers, and only then did email gradually recover.

Microsoft apologised to its customers, extended the retry time for email to forty-eight hours, and added capacity and monitoring.

This reminded many people of the CrowdStrike update accident of 2024, when one small error likewise brought computers around the world to a stop.

The full investigation report, Microsoft says, will be published in a few days.

Listen again

Try it without the transcript and notice what sounds clearer.

What vocabulary does this episode teach?

词汇
bànfǎway; method

HSK 2. Used in 修电脑最老的办法.

jiéguǒas a result; result

HSK 2. Used in 结果越修越糟.

yuányīncause; reason

HSK 2. Used in 真正的原因.

yālìpressure; load

HSK 3. Used in 这些服务器的压力一下子大了.

zēngjiāto increase; to add

HSK 3. Used in 增加了容量和监控.

huīfùto recover; to restore

HSK 5. Used in 邮件才陆续恢复.

chéngrènto admit

HSK 4. Used in Microsoft 还承认.

liúliàngtraffic (data flow)

The volume of data moving through a network.

wǎngguāngateway

The servers that connections pass through on the way into a service.

liánsuǒ fǎnyìngchain reaction

One failure causing the next, like links in a chain.

jīyāto pile up; backlog

积压的流量 is the traffic that built up while the servers were down.

chùfāto trigger

触发了那个漏洞 means it set off that flaw.

* beyond level超纲词

What grammar patterns appear in this episode?

语法

越 A 越 B

The more A, the more B.

关掉再打开,是修电脑最老的办法,十一月 Microsoft 的工程师也试了,结果越修越糟。

Microsoft 还承认,平时的负载测试没有测出这个漏洞,而且机器越新、核心越多,问题越严重。

被 + Verb↗

The passive: the subject receives the action.

一台服务器失去反应,流量被转给别的服务器,别的服务器也跟着倒下。

…… 才 ……

才 shows that something happened only then, later than hoped.

十三号,工程师给服务器装上修复补丁,邮件才陆续恢复。

也就是说↗

In other words. Restates the point more plainly.

Microsoft 很早就表示没有发现恶意活动,也就是说,它不认为这是一次攻击。

Proper nouns

专有名词
美洲Měizhōuthe Americas

Sources

来源

Free account

Keep learning from this story

Create a free account to keep saved words and your preferred level together.

  • Keep words with their story context
  • Remember your preferred level
  • Build your vocabulary over time

Keep learning