Part two of the batch tuning series delves into practical analysis using observability tools, offering insights and recommendations for tuning batch sizes in Redpanda. The focus is on understanding the impact of effective batch size on system performance, including efficiency improvements, resource utilization, and latency reductions. Metrics such as batch size, CPU utilization, and scheduler backlog are analyzed using Prometheus and Grafana to provide a comprehensive view of the system's behavior. A real-world example demonstrates how optimizing batch sizes led to significant performance gains, reduced CPU usage, and network bandwidth savings for a customer migrating from Kafka to Redpanda Cloud. By systematically adjusting configurations like linger time and batch size, substantial improvements in latency and resource efficiency were achieved, illustrating the importance of batch tuning for maximizing throughput and minimizing costs. The discussion concludes with the potential for further enhancements by consolidating data flow to fewer clusters due to the increased capacity achieved through effective batch size tuning.