I am getting occasional disappearing of data points on dashboard. I have used prune scripts in the smartstor to prune aged historical agents and also moved collectors to a cluster on ESX host where there is less contention of resource to no avail.
This is what I my setting is across all collectors and MOM
# Threshold for when a collector is considered slow (in milliseconds)
# Seconds a MOM waits for a Collector to respond before it disconnects it as hung.
# Default is 60 seconds. Setting to zero (or negative) disables the check (not recommended).
# The MOM checks collectors every 7.5 seconds, so small changes don't affect behavior.