Salesforce's Cloud Outage Exposes Hidden Risks of Complex Dependencies
On September 16th, Salesforce experienced a massive service outage that lasted for nearly seven and a half hours, impacting multiple instances across all regions. The incident occurred at 3:50 a.m. EDT and was marked resolved around 3 p.m. EDT.
Salesforce attributed the issue to an 'external dependency failure' affecting the legacy login server, which experienced increased load due to its capacity to process requests. Despite blocking API endpoints and attempting rolling restarts, some automated fixes didn't fully resolve the problem.
According to Info-Tech Research Group's principal advisory director Abbas Jaffery, the incident highlights a cloud resilience issue rather than just a legacy one. He emphasized that cloud does not eliminate architectural dependencies, which can become an enterprise risk when the platform is a system of record.