In a recent help ticket, we found that the logs and reporting in the shard registry was insufficient to easily identify an inconsistency between the read through cache's time in store and the local topology time. Some things that could benefit from additional debug abilities are:

1. Provide a way to dump the read through cache's state, especially the time in store (could also be useful for other instances of the read through cache as well as the shard registry)
2. Add debug logging when advanceTimeInStore does not actually cause an advance in the read through cache and the refresh was intended as a force