We have retry logic in place for situations with no network connection. Events (logs/moments/crashes) are added to a queue and when the connection is restored, we attempt to gather the events.