[clappski]

There's lots of individual data points in time series data, so having a high file:data point ratio would cause issues including running out of inodes, the identifying handle of a file at the OS level.