The September 16, 2026 Salesforce Outage: What the Hyperforce Incident During Dreamforce Reveals

On September 16, 2026, in the middle of Dreamforce, Salesforce suffered a major outage affecting Hyperforce instances across all regions. According to the official report published on status.salesforce.com (Salesforce Trust), the incident started at 07:50 UTC and was declared resolved at 15:26 UTC — roughly 7.5 hours of disruption before service returned to normal — with the incident formally closed, after extended validation, at 18:59 UTC. Here is exactly what the official report states, without repeating tech-media speculation about the precise cause.

Incident timeline (source: Salesforce Trust)

• 07:50 UTC: service disruption begins (official impact timestamp)
• 08:32 UTC: Salesforce publicly confirms it is investigating an issue affecting multiple instances across all regions, with severe delays, intermittent errors and inability to access some services; support case creation also affected
• 09:10 UTC: investigation points to requests stalling while waiting for a response from an internal login service, consuming available server resources
• 09:57 UTC: Salesforce cites an external dependency failure impacting the legacy login server, and rules out third-party infrastructure after checking with its provider
• 10:56 UTC: the fix is validated on a test instance
• 11:19 – 14:56 UTC: progressive rollout of the fix region by region, with manual restarts on instances where the automated fix didn't fully resolve the issue
• 15:26 UTC: no further impact confirmed on the affected Hyperforce environments
• 18:59 UTC: incident formally closed after extended telemetry validation

Technical cause per the official report

The Salesforce Trust report describes two contributing factors: a core system component experiencing increased load that limited its capacity to process requests, and an external dependency failure impacting the legacy login server. Salesforce explicitly states that a full root-cause analysis is not yet published as of the incident's closure: the company commits to "establishing the technical trigger, the underlying cause, and preventive action" in a forthcoming investigation. Since no date is given for that detailed report, we are not speculating on its content.

Who was affected

• Hyperforce instances across Salesforce's three operating regions (isCore: true, affectsAll: true in the official report) — a subset of instances was eventually identified as the real scope, narrower than the initially estimated impact
• "First-party" environments (legacy, non-Hyperforce infrastructure) were not impacted, per Salesforce
• GovCloud customers were confirmed out of impact
• Some customers were unable to create new support cases through the Help portal during the incident
• After recovery, some customers reported scheduled jobs not running as expected — a point Salesforce says it investigated separately

What this means for your business

• A 7.5-hour outage on one core component (a login service) shows that an incident on a single technical element can cascade across every region of a major SaaS vendor — check whether your organization depends on critical automations (scheduled jobs, synchronous integrations) that assume continuous Salesforce availability
• If your business relies on creating support cases through the Help portal, keep a documented fallback channel (direct contact with your Account Executive or TAM) for incidents where that channel itself is unavailable
• After an incident of this scale, verify your scheduled jobs and integrations actually ran during and after the impact window (07:50–15:26 UTC on September 16, 2026) rather than assuming a full automatic catch-up
• Since the root-cause report isn't public yet, don't draw a final conclusion about the exact cause before it's published — follow status.salesforce.com for the official update rather than a third-party summary

Want to assess how resilient your Salesforce architecture actually is against this kind of incident? Our consulting offering audits your critical dependencies (scheduled jobs, synchronous integrations, support channels) to identify weak points before an incident reveals them for you.