Updated Saturday 11:06am PT: Salesforce sounded the all-clear Saturday morning, ending an outage that lasted 15 hours and 8 minutes according to its system status page. Our original story follows below.
A database script error forced Salesforce engineers to cut service to a wide swath of customers Friday, and for some customers using its Pardot service, Salesforce has been down all day.
Just before 10am PT Friday morning, Salesforce engineers deployed a database script that “inadvertently gave users broader data access than intended,” it said on its status page. That is not good, and while the broader access issue only affected customers who currently use Pardot or have used it in the past, in order to prevent further problems the company decided to cut service to all customers using instances that hosted Pardot users.
As of 2:49pm PT, Salesforce said that it had restored access for customers who have never used Pardot, a “customer success” tool that helps marketing teams sell you more stuff. However, current and former Pardot customers remain unable to access the core Salesforce service as of this writing, and the company is “working to restore the original permissions as quickly as possible for customers that were affected by the permissions change,” it said.
Salesforce also said a “subset” of customers were affected by the issues, but dozens of Salesforce instances across North America and Europe were listed on the outage notice.
A Salesforce representative did not immediately reply to a request for more details on the outage and an expected time of resolution.
Needless to say, that’s a long outage on a late spring Friday, and there are lots of sales teams that didn’t get much done today. We’ll update this post as more information becomes available, but in the meantime, here’s an artisanally curated number of outage tweets:
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. Please know that we have all hands on this issue and are resolving as quickly as possible.
— Parker Harris (@parkerharris) May 17, 2019
— Andrew Ray (@andrewray) May 17, 2019