Previous incidents

Postmortem is available at here.

The AU auth service has recovered. Azure completed the database update and traffic is stable again.

The issue came from an internal component used by our proxy, which caused incorrect status reporting. We're updating our error handling to prevent this going forward. Service availability was not affected.

The abnormal traffic source was contained and affected domains were temporarily blocked. Performance returned to normal for all tenants. We’re contacting the customer involved to verify intent and ensure proper setup going forward.

Azure has mitigated the underlying incident affecting services in West US 3. Our monitoring shows that error rates have returned to normal. If you still see any issues, please let us know.

Our components were healthy throughout. Upstream infrastructure logs showed 5xx responses and timeouts, while our own service logs showed no corresponding errors. This indicates an issue in the underlying platform. We’re continuing to confirm.

All authentication services are now fully restored. Traffic is back to normal. We’ll keep monitoring.

The incident was traced to transient network jitter from our cloud provider. Service performance has recovered, and we’ll continue to monitor for any recurrence.