The design is good but has potential single points of failure, such as the main database and CDC connector. You can address this by using high availability and failover configurations for the database (e.g., replication and clustering) and implementing monitoring, alerting, and redundancy for CDC connectors