Take into account slab cache in Memory Utilization

Following up on issue PMM-9513 I propose to add the slab cache metric to the "Cached" value in Node Summary -> Memory Utilization. Under certain circumstances (i.e. if the node uses a large slab cache) the "Cached" value may be way too low, and therefore the "Used" value too high, causing an inconsistency between the Memory Utilization and Virtual Memory Utiliztation chart.