Loading live status…
Loading live status…
On 16 September 2026, the second day of Dreamforce, Salesforce had a service disruption across instances in every region, tied to an internal login service. Here is the timeline, what Salesforce said caused it, and what to do before the next one.
On 16 September 2026, Salesforce had a major service disruption. It hit multiple instances in all regions, and customers from the US to Japan, India, the UK, France and Germany reported problems. It also happened on the second day of Dreamforce, while Salesforce was on stage in San Francisco pitching AI agents that run businesses.
This post pieces together what is publicly known from Salesforce's status updates, as reported by several outlets. Reported times differ slightly between sources, so we give only the milestones that line up.
| Time | What happened |
|---|---|
| 12:50 AM PDT (07:50 UTC) | Disruption begins across multiple instances in all regions. Users see severe delays, intermittent errors and failed logins. |
| Early morning PDT | Salesforce confirms widespread impact. It says requests are stalling while waiting on an internal login service. |
| About 3:30–4:00 AM PDT | Salesforce points to "increased load on a core system component", validates a fix on a test instance, and begins rolling it out region by region. |
| Late morning ET | According to CIO.com, most customers are back online and mitigations are in place across all instances. |
| Afternoon ET | The incident is marked resolved. |
Some Hyperforce instances needed manual restarts and recovery work after the automated fix did not fully take, The Register reported. Salesforce said customer data was not affected, and the problem was access, not data loss.
Salesforce's public explanation came out in steps during the incident:
Salesforce said: "We will undertake a full investigation of the incident, establishing the technical trigger, the underlying cause, and preventive action to avoid a repeat in the future." Until that analysis is published, treat any single-sentence root cause, including the ones above, as provisional.
Because the fault was in the login path, the impact went well beyond people who couldn't open the CRM. Any system that signs in to Salesforce before doing its work was affected. That includes middleware syncs, marketing tools, data pipelines and, increasingly, AI agents.
The same week, Salesforce was making the case for agents that work for days or weeks and for putting Slack and Claude in front of CRM data. The outage is a reminder that all of that depends on one thing at the bottom of the stack: being able to authenticate. The more workflows you move onto a platform, the more a login-layer failure costs you.
It is also a reminder that "all regions" failures happen even on multi-region cloud platforms. Hyperforce spreads Salesforce across public cloud regions, but a shared internal service can still be a single point of failure across all of them.
Sources
Checked continuously against each provider's own status feed.