What Is High-Cardinality in Monitoring?

Question
High-cardinality in SQL refers to highly differentiated data for a given column. (Although "relatively unique" is not grammatical, that is one way to say it.) When there are many different possible values for a given column, that is high-cardinality. In the context of monitoring, what is high-cardinality?

Answer
In monitoring like relational databases, cardinality refers to the number of elements in a set (consistent with the mathematical definition https://www.dictionary.com/browse/cardinality).

"[T]he number of individual values of a metric." This was taken from Splunk.com.

Time series data can be paired with labels, meta-data or dimensions. This combination increases cardinality from the atomic metric categories.

Creating artificially high cardinality can create large amounts of data to reconcile; some businesses demand high cardinality for certain analytical purposes.

See also this posting: https://blog.timescale.com/blog/what-is-high-cardinality-how-do-time-series-databases-influxdb-timescaledb-compare/

Leave a comment

Your email address will not be published. Required fields are marked *