Overview
- A widespread service disruption began on Wednesday, September 16, when customer requests stalled waiting for a response from a legacy internal login service, which consumed server capacity and caused delays, intermittent errors and login failures across hundreds of instances worldwide.
- Investigators traced the problem to an internal authentication bottleneck linked to an external dependency, and Salesforce blocked an implicated API endpoint while engaging a third-party infrastructure provider for diagnostics.
- Engineers validated a code fix in a test environment early that morning and began a fleetwide rollout by mid‑morning Eastern, after which Downdetector reports and company status updates showed complaint volumes falling as most customers saw services return to normal.
- The outage blocked some customers from creating support cases and disrupted daily workflows for large clients, producing a modest premarket stock dip and heightened scrutiny because it occurred during the company’s Dreamforce conference.
- Salesforce says residual delays persist for some instances while teams work on a permanent, code-level solution; the incident highlights how legacy authentication systems and external dependencies can cascade into broad outages for large cloud platforms.