Salesforce went down worldwide during its own Dreamforce. Full recovery took nearly 11 hours.
The Register said the disruption was first reported around 09:30 BST on 16 September.
Instances were hit in the USA, Japan, India, the UK, France and Germany. Salesforce's status page warned of "severe delays, intermittent errors, and inability to access some services". Customers could not open support cases either.
Salesforce said "requests are stalling while waiting on a response from an internal login service". CIO reported that the company later traced it to "an external dependency failure impacting the legacy login server".
A fix was validated and rolled out at 12:19 BST, region by region. According to The Register, some instances were still not restored at 15:00. Some customers needed manual restarts, while scheduled jobs failed to run for some users. Salesforce declared the incident resolved at 20:20 BST, nearly 11 hours after the first reports.
Dreamforce expected more than 40,000 people in person in San Francisco that week.
Abbas Jaffery of Info-Tech Research Group told CIO: "Cloud does not eliminate architectural dependencies. It can sometimes make them less visible."
Which of your integrations would have failed silently when the login did?
Sources
Our file on Salesforce
- 20 Sept
Salesforce set up a committee. It picks which AI model runs a task.
- 18 Sept
Salesforce says AI cut Air India's refunds to four hours. From 14 days.
Every story here is open to read. The ERP LEADERS brief goes one step further.
One ERP programme per issue, laid out for a steering committee. Issue 01 is the Zeiss case. Read issue 01 or sign up for the brief.
Welcome back. · Issue 01