A bad mistake by GoDaddy took Zoom offline for almost two hours on Wednesday afternoon, US time.
Zoom explained the situation in an incident report that opened with a 12:17 PDT (Pacific Time) update that advised customers “We are investigating domain name resolution issues on the zoom.us domain that is affecting multiple services.”
Zoom later admitted the outage started at 11:25 PDT.
We understand that video meetings hosted on Zoom were interrupted mid-stream, with “This site can’t be reached” or “Check if there’s a typo in zoom.us” errors appearing on-screen.
Users who tried to reach the Zoom status page at status.zoom.us were out of luck – it was down too. Zoom account managers were hard to reach as most use Zoom’s VoIP phones to communicate with customers.
Cisco’s ThousandEyes observability outfit analyzed the incident and picked it as a DNS problem that meant top-level domain nameservers did not have the records for zoom.us.
“The issue had a cascading effect that impacted Zoom’s services, particularly their main webpage, zoom.com,” ThousandEyes added. “This indicates that the content delivery network (CDN) serving Zoom was unable to connect to the backend services hosted on zoom.us.
That assessment was correct, but we’ll get to that in a bit.
Service returned at 13:55 PDT and resuming Zooming required users to use command line skills to flush their DNS caches.
IT departments, take note: Zoom’s instructions for restoring services on Macs instructed users to type sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder, which means plenty of your user populations now have an inkling about the power of sudo
The final incident report update, time-stamped 17:31 PDT, revealed the cause of the incident:
“On April 16, between 2:25 P.M. ET and 4:12 P.M. ET, the domain zoom.us was not available due to a server block by GoDaddy Registry. This block was the result of a communication error between Zoom’s domain registrar, Markmonitor, and GoDaddy Registry, which resulted in GoDaddy Registry mistakenly shutting down zoom.us domain.”
Markmonitor is a domain management and security outfit. GoDaddy Registry manages the entire .us namespace.
If its stewardship of .us domains can see it take them offline, the org has a fair bit of explaining to do.
The final entry in the incident report continues:
“Zoom, Markmonitor and GoDaddy worked quickly to identify and remove the block, which restored service to the domain zoom.us. There was no product, security, network failure or Distributed Denial of Service (DDoS) attack at Zoom during the outage. GoDaddy and Markmonitor are working together to prevent this from happening again.”
The Zoom outage took place at the same time as music-streaming Spotify wobbled for some users, but the two incidents appear to be unrelated. ®
