Hello,
You will see this error occasionally. When the main instance is restarted, connections might be routed to your replica and you will get this message for write requests. We recommend that you implement error handling strategy like exponential backoff to reduce the amount of errors when this happens.