Vercel outage on 18 November 2026
An email when it goes down, and when it’s back.
“Delays in observability data”
On 18 November, Vercel had a major outage, opened on its status page at 20:44 UTC and resolved at 23:53 UTC (about 3 hours 9 minutes). The status page said observability and billing data were delayed, Workflow runs could show as pending after they had started or finished, and AI Gateway observability could show low traffic. AI Gateway, Workflow, Observability and, from 20:55, Queues were affected. A fix was reported in place at 21:30 with recovery under way, though the incident stayed open until 23:53. Musixmatch, Auto Trader UK, SurveyMonkey, Apple, Aircall and Societe.com also reported problems during the same period.
Vercel’s own incident page ↗- When
- 20:45–23:54 UTC
- How long
- 3 h 9 min
- Incident open on its page
- 20:44–23:53 UTC
- Components affected
- 4
- Vercel, 30 days before
- No other outages
What broke when
From Vercel’s own status page: when it moved each component into trouble, and back.
At the same time
Started in the hour before or during it
And 153 lesser-known services. Problems at the same time are a clue, never proof that one caused another.
Timeline
Times in UTC
- Investigating
- AI Gateway: partial outage
- Workflow: partial outage
- Observability - Observability: partial outage
- Cause identified
- Queues: partial outage
- Fix in place, monitoring
- Resolved
- AI Gateway: back to normal
- Workflow: back to normal
- Observability - Observability: back to normal
- Queues: back to normal
Where this comes from
Stackpulse reads Vercel’s official status page at www.vercel-status.com every minute. The times above are when it reported the problem; the summary is written from its updates and checked against them.
Stackpulse isn’t affiliated with Vercel. Names and logos belong to their owners.