Time-series databases are optimized for performance based on having an index for each label/type of value. That inversion allows for a tremendous amount of valu

Time-series databases are optimized for performance based on having an index for each label/type of value. That inversion allows for a tremendous amount of values to be captured and reported against with low latency, but with the trade off of having less permutations available for scale. The same applies to Prometheus, Graphite, etc.