Salesforce outage: What happened as customers faced delays and errors worldwide

Salesforce customers around the world faced a major service outage on Wednesday, with users reporting severe delays, intermittent errors and problems accessing some Salesforce services. The outage affected hundreds of Salesforce instances across several countries, including the US, Japan, India, the UK, France and Germany.Salesforce services faced a major outage affecting customers worldwide. (AP Photo/Ted Shaffrey, File) (AP)The service disruption was first reported at around 9:30 BST (8:30 UTC) on September 16, according to Salesforce's status page. Salesforce initially said its investigation found that customer requests were getting stuck while waiting for a response from an internal login service. The company said this was using up available server resources.Salesforce outageThe company said the issue was serious enough to affect support case creation as well, meaning some customers were unable to create cases with Salesforce support during the outage. The company said customers were still facing severe delays, intermittent errors and difficulty accessing some services.At 11:42 BST (10:42 UTC), Salesforce gave more details about the problem. It said increased load on a core system component had reduced its ability to process customer requests.Salesforce outage fixThe company said it would deploy the fix across its wider system once testing showed that the solution was working. It then started preparing to deploy the fix across all affected instances.By 12:27 BST (11:27 UTC), Salesforce said the fix was being rolled out across its entire affected fleet. The company also said customers were beginning to see services return to normal.The outage affected Salesforce at a particularly important time for the company because its annual Dreamforce conference began in San Francisco on Tuesday.Therefore, the key issue identified by Salesforce was linked to increased load on a core system component and requests waiting for an internal login service, which affected the system's available capacity to handle requests.