When running multiple Region Servers on a single node monitoring tools are unable to distinguish between the metrics, because they group them by hostname. E.g. Ganglia shows accumulated metrics for "hbase.regionserver.blockCacheCount" for all Region Servers running on the same node.
If a Region Server ID was added to the metric, e.g. "hbase.regionserver001.blockCacheCount", then Ganglia will be able to display the metrics for each Region Server separately.