It can be difficult to know if a Policy Engine has failed to connect to its parent for replication, particularly in large environments with 250+ PE servers. In this case, even the performance counters for the replication cache may fail to be updated. It would be helpful if the dump of the replication cache could be configured to occur on a regular basis, instead of just at startup/shutdown or by manual command. This would allow an easy way to monitor whether the cache is building up due to a replication issue. Ideally, the option should be configurable via machine policy.