Sign-in and authentication failures in Automation Cloud - All regions
Final update
## Customer impact Between November 18, 2026 at 2:00 pm UTC and approximately 2:45 pm UTC, a subset of customers experienced intermittent failures signing in to Automation Cloud and requesting access tokens. Impact was limited to organizations whose authentication requests were routed through the affected North Europe component; for those organizations, requests to Orchestrator and other Automation Cloud services that depend on access tokens failed intermittently regardless of which region they were working in. Customer-facing impact lasted approximately 45 minutes. ## Root cause The incident was caused by a deployment of the authentication service that replaced all active instances in the North Europe region at once. The new instances received full production traffic while still warming up and could not complete startup within the service's five-second health-check timeout, so they were repeatedly restarted. Because no instance ever reported healthy, automatic scaling could not add capacity to relieve the load. The simultaneous replacement and the undersized health-check timeout were both required for the failure; neither alone would have caused customer impact. ## Detection Autom
Timeline
- Postmortem · Nov 20, 23:55 UTC
## Customer impact Between November 18, 2026 at 2:00 pm UTC and approximately 2:45 pm UTC, a subset of customers experienced intermittent failures signing in to Automation Cloud and requesting access tokens. Impact was limited to organizations whose authentication requests were routed through the affected North Europe component; for those organizations, requests to Orchestrator and other Automation Cloud services that depend on access tokens failed intermittently regardless of which region they were working in. Customer-facing impact lasted approximately 45 minutes. ## Root cause The incident was caused by a deployment of the authentication service that replaced all active instances in the North Europe region at once. The new instances received full production traffic while still warming up and could not complete startup within the service's five-second health-check timeout, so they were repeatedly restarted. Because no instance ever reported healthy, automatic scaling could not add capacity to relieve the load. The simultaneous replacement and the undersized health-check timeout were both required for the failure; neither alone would have caused customer impact. ## Detection Automated gateway health alerts triggered the incident at 2:17 pm UTC on November 18, 2026, approximately 17 minutes after the first observed customer impact began. The alert was followed by multiple service teams reporting authentication-dependent failures. ## Response At 2:10 pm UTC, automated traffic
- Resolved · Nov 18, 15:26 UTC
The authentication issue has been fully resolved. Customers should now be able to sign in to Automation Cloud and request access tokens successfully. Orchestrator and other Automation Cloud services are operating normally. Our monitoring confirms that system health has returned to normal, and no further impact has been observed. The incident is now resolved. Between 14:00 and 14:45 UTC, a deployment issue affected one of our Identity service cluster in Europe. Because authentication requests are routed according to the customer organization’s assigned cluster, some customers in multiple regions experienced intermittent sign-in and access-token failures, which could also cause requests to Orchestrator and other Automation Cloud services to fail We will provide a detailed root-cause analysis (RCA) as soon as it is available
- Monitoring · Nov 18, 15:02 UTC
The authentication issue has been mitigated, and our monitoring indicates that overall system health is recovering. Our engineering team continues to closely monitor the environment and validate service performance to ensure sustained stability. We will provide additional updates as the situation progresses.
- Investigating · Nov 18, 14:36 UTC
Starting at 14:00 UTC on November 18, some customers may experience intermittent failures when signing in to Automation Cloud or when services and robots request access tokens. As a result, requests to Orchestrator and other Automation Cloud services may fail intermittently with errors. Our engineering team has identified the affected authentication component and is actively working to restore service.
- Investigating · Nov 18, 14:35 UTC
Starting at 14:00 UTC on November 18, some customers may experience intermittent failures when signing in to Automation Cloud or when services and robots request access tokens. As a result, requests to Orchestrator and other Automation Cloud services may fail intermittently with errors. Our engineering team has identified the affected authentication component and is actively working to restore service.