The outage began on Friday after the deployment of a database script that inadvertently granted users broad permissions over companies’ data
The incident began at about 6 p.m. BST on Friday after Salesforce deployed a database script that resulted in granting users broader access to firms’ data than intended, Salesforce said.
Salesforce users said on social media that the update had given all users full access to read and modify all of a company’s files.
The permissions issue only directly affected users of Pardot, a marketing automation tool purchased in 2012, but out of an abundance of caution Salesforce decided to completely shut down access to more than 100 cloud instances that host Pardot users.
The company said on Saturday that it had restored access to all firms that hadn’t been directly affected by the permissions issue.
In an update on Monday morning, Salesforce said it had executed automated provisioning to restore permissions on all production instances, but that some customers may still have been experiencing permissions issues.
A status page continued to list disruption to dozens of instances.
“To all of our Salesforce customers, please be aware that we are experiencing a major issue with our service and apologize for the impact it is having on you,” said Salesforce CTO and co-founder Patrick Harris in a message on Twitter on Friday. “Please know that we have all hands on this issue and are resolving as quickly as possible.”
The glitch mainly affected users in Europe and North America, according to Salesforce.
The company said it was providing administrators who continued to experience permissions issues with instructions on manually restoring them.