Salesforce Instance Failure Causing Widespread Service Disruption

A power outage seems to have affected networking gear, bringing down one of nearly 200 Salesforce instances powering North American customers. The CRM giant is rerouting to an emergency site.

Many Salesforce customers lost access to the company’s service Tuesday when one of the cloud software giant’s nearly 200 virtual servers in North America went down due to an apparent power outage.

The service disruption that started at 11:54 am Pacific Time seems to only have affected users hosted on Salesforce’s NA89 instance, according to Salesforce’s status page and notifications sent to customers.

The North America 89 instance runs in data centers in Phoenix and Washington D.C., with transactions replicated across those locations for redundancy.

“Upon investigation, the Technology team believes this may be caused by a power outage hosting the network switches responsible for routing NA89 traffic,” Salesforce said.

[Related: Salesforce CEO Marc Benioff Reportedly Hints At No Digital Dreamforce This Year]

Salesforce is currently executing an emergency site switch, which it began working on at 12:15 pm Pacific Time. The San Francisco-based company said it would provide updates soon.

Downdetector.com, which tracks cloud outages, saw a spike to more than 300 problems reported by Salesforce customers starting just before noon PT.

In May of last year, Salesforce saw a major outage triggered by a Pardot marketing automation software scripting error. Although the scripting error only affected the Pardot service, Salesforce initially blocked access to all instances of its software for affected organizations.

A more significant outage that lasted almost a day in May of 2016 resulted in some customers losing four hours of data.