Every secure connection on the web leans on a certificate: a signed file that proves a server is who it says it is. Certificates are issued with an end date. After that date, software stops trusting them, and connections that depend on them fail.
On Monday 3 February 2020, Microsoft Teams went down from about 8:30 a.m. Eastern, TechCrunch reported. People signing in, SiliconANGLE reported, got errors saying the app had failed to make a secure HTTPS connection to Microsoft's servers.
Microsoft's status account posted three updates, about an hour apart by the press's timing:
- It was investigating "an issue where users may be unable to access Microsoft Teams".
- Then the cause: "We've determined that an authentication certificate has expired causing, users to have issues using the service. We're developing a fix to apply a new certificate to the service which will remediate impact."
- Then: "We've initiated the deployment of the updated certificate".
Users soon reported Teams working again. SiliconANGLE put the outage at about three hours. TechCrunch noted Teams had about 20 million users at the time, and that the outage came in the middle of Microsoft's fight with Slack for business chat.
The sting, as SiliconANGLE pointed out, is that Microsoft also builds and sells tools meant to monitor things like expiring certificates. An expiry date is a failure you can see coming long before it happens. It just needs someone, or something, to be watching the calendar.
- About 8:30 a.m. Eastern, 3 February 2020: Teams goes down (TechCrunch)
- 3 status updates, about an hour apart
- About 3 hours offline, by SiliconANGLE's account
- About 20 million Teams users at the time, per TechCrunch
Why it matters: Certificates fail on a date you already know. Keep an inventory of every one, alert weeks before expiry, and automate renewal so no outage depends on someone remembering.