At 1:39PM PDT we were alerted to an issue where metrics were not syncing. We restarted a syncing service and processed the backlog of labels. We installed a failover second syncing service, and are restructuring our syncing infrastructure to better handle errors and congestion.