Salesforce suffered a widespread outage on Wednesday, September 16, leaving customers across its global infrastructure facing severe delays, intermittent errors and an inability to access parts of the cloud platform.

The disruption began at approximately 7:50 a.m. UTC, or 3:50 a.m. Eastern Time, during the second day of Dreamforce 2026 in San Francisco. The timing placed the failure in full view of customers, developers and technology executives attending Salesforce’s biggest annual event.

Salesforce reported problems across multiple instances in all operating regions. Some customers were also unable to create support cases through the Help portal, limiting a crucial escalation route during the incident.

Login requests placed pressure on core systems

The first technical findings pointed to a bottleneck inside the login path. Salesforce said requests were stalling while waiting for a response from an internal login service. Those unfinished requests remained active and consumed available server resources, leaving the platform with less capacity to process other traffic.

The effects extended beyond people attempting to sign in through a browser. Mobile sessions and API integrations also rely on authentication services before they can retrieve records or complete transactions. A slowdown in that shared pathway can therefore produce login failures, integration alarms and timeouts across otherwise separate business applications.

Engineers initially attempted a rolling restart on one affected instance. The company later abandoned restarts as its main recovery route after determining that increased load on a core system component was limiting request-processing capacity.

Salesforce also investigated a possible external dependency failure affecting a legacy login server. Engineers considered blocking an API endpoint as a temporary mitigation and contacted the relevant third-party infrastructure provider. The provider reported no corresponding problem within its own infrastructure, leaving the precise original trigger unconfirmed.

By 3:31 a.m. Pacific Time, engineers were testing a targeted fix on a controlled instance. Salesforce validated that fix at approximately 3:56 a.m. and then began deploying it across the affected production fleet.

Salesforce reports improvement as rollout continues

In its latest public statement, Salesforce said: “The rollout is progressing region by region and we’re seeing improvement. We don’t have a confirmed time of completion.”

The company confirmed that GovCloud customers were no longer affected, although other customers could continue to experience delays, errors and failed access while the regional deployment continued. Salesforce said it was also developing a permanent code-level fix alongside the immediate recovery measure.

The available evidence does not yet amount to a final root-cause finding. It shows that login-related requests became stuck, consumed server resources and increased pressure on a core component. Salesforce has not yet identified the exact defect or event that started the sequence.

The incident was separate from a user-interface freezing issue affecting some Chrome and Edge 153 users. Salesforce connected that earlier problem to an external Chromium update, while Wednesday’s outage involved server-side request processing across browsers, APIs and regions.

For businesses running sales, customer service and automated workflows on Salesforce, recovery may involve more than restoring access. Timed-out API transactions, delayed automations and middleware retry queues may still require reconciliation once the platform stabilises.

Salesforce shares moved only slightly in early trading, falling around 0.3% to 0.4%. The limited reaction suggested investors initially viewed the outage as a contained operational problem, despite its uncomfortable arrival during Dreamforce.

The company’s official Salesforce incident record remains the primary source for regional recovery updates and any eventual root-cause report.