I have the following stats on a 4 node cluster, I am wondering why it is only using 84% of the ram allocated to the bucket instead of keeping more items in memory. Is metadata overhead not reflected in the free RAM “Amount of RAM available on this server” metric for the bucket?
453,901,466 item count
ram / quota: 347GB / 410GB
data / disk: 426GB / 563GB
It lists 73 GB free RAM but only a 61.2% active document resident %
Are there any settings to change to make it use another 60 gigs of ram? It seems like it is just letting that ram sit there unused. While loading the free ram steadily matched the item count in the bucket, but now free ram has stopped decreasing and is staying at 73 GB