The incident was caused by a previously identified operating system–level issue. This issue was already under investigation, and corrective OS changes had been scheduled as part of a planned maintenance window. Before the scheduled maintenance could be carried out, the issue manifested and resulted in a brief service disruption, during which all active connections were dropped for approximately one minute. As this behavior was consistent with the known issue, it was anticipated. The scheduled maintenance (details published in advance on our status page) is intended to apply the required fixes and prevent further occurrences of this problem.
