There was an upgrade in the cloud infrastructured which triggered a restart of the service. This caused about 5 minutes of downtime.