Customer impact

Between May 13, 2026, 1:50 pm UTC and 2:25 pm UTC, customers using Automation Cloud services in our Switzerland region experienced service unavailability. Failures in the licensing service prevented licensing checks from completing, which in turn made parts of Automation Cloud inaccessible. Impact was limited to the Switzerland region and lasted approximately 35 minutes. Services began recovering at 2:16 pm UTC, with full resolution confirmed at 2:25 pm UTC.

Root cause

The issue was caused by a deployment configuration change in the licensing service that did not apply cleanly. The change was part of a planned resource rename intended to streamline future licensing deployments. A prior deployment earlier that week did not complete successfully, leaving the rename in an intermediate state with no immediate warning that something was wrong. The current deployment in the affected environment still references the old resource name. At the start of the incident, the licensing service needed to access the resource, which caused dependent service calls to fail and surfaced to customers as service unavailability.

Detection

Automated monitoring detected the issue at 1:50 pm UTC when multiple health checks for licensing-dependent workflows in Switzerland failed simultaneously. The incident was declared immediately upon alert firing, with no measurable gap between detection and declaration. An initial investigation quickly localized the failure to the regional licensing service rather than its dependents.

Response

The incident was opened at 1:50 pm UTC and responders joined the incident’s bridge immediately. By 2:04 pm UTC, the team had identified the cause, prepared mitigation steps, and updated the public status page to indicate that restoration work was underway.

Mitigation involved correcting the deployment configuration in the affected environment and retrying the deployment. Health checks began recovering shortly after, and service recovery was confirmed at approximately 2:12 pm UTC while the deployment continued to roll out across the region.

At 2:16 pm UTC, the incident was marked as mitigated and the status page moved to monitoring. At 2:25 pm UTC, the deployment completed successfully and the incident was marked as resolved.

Follow-up

We are evaluating preventive measures to ensure that incomplete deployments do not leave services in an inconsistent configuration state. The team is prioritizing improvements to deployment validation and to the detection of configuration drift between environments, so that mismatches like this are caught before they reach production traffic.