pulsar bookie ledger count

3节点的集群：zk+bookie+broker

通过图上所示，监控上没有发现backlog堆积
问题1：通过命令行获取的ledger数量和监控取值的，两个差别怎么如此巨大？
问题2：如果ledger的数量是图上listledgers获取的值所示，那么bookie后台还在清理过期消息，为啥还有那么多的ledger呢？请问导致这种大的ledger是什么情况会触发【例如：大量写入消息】？

Start by comparing the ledger count returned by the command-line listledgers operation with the monitoring value for the three-node ZooKeeper, bookie, and broker cluster. Investigate how expired-message cleanup relates to the reported ledger total and identify the conditions that produce many ledgers. Done means explaining the discrepancy and the apparent accumulation without relying on backlog growth.

Written by the indexing model from the issue text.