This gauge reports the average amount of cached memory over the sampled interval on the fabric-interconnect switch's onboard OS, as recorded in the currently sampled CPU-load and memory-usage statistics for the fabric-interconnect switch's onboard OS. An administrator distinguishing genuine memory pressure from harmless caching behavior on the switch supervisor compares this memCachedAvg figure against memAvailable, since cached memory can normally be reclaimed on demand. For example, a large memCachedAvg figure alongside healthy memAvailable would tell an operator the switch's memory usage is benign filesystem caching rather than a sign of trouble.