14 Managing cardinality in telemetry


This chapter covers

  • How cardinality affects telemetry performance
  • Ways to identify cardinality problems
  • Techniques for managing cardinality

This chapter dives deep into one of the maintenance headaches of managing telemetry systems: cardinality in your storage systems. Here is the definition of cardinality from chapter 1:


Cardinality is the term for index complexity—specifically, the number of unique combinations the fields in the index may produce. If you have fields A and B, where A has two possible values and B has three possible values, the cardinality of that index is A * B, or 2 * 3 = 6. Cardinality significantly affects search performance no matter what data storage system is being used.

14.1 Identifying cardinality problems

14.1.1 Cardinality in time-series databases

14.1.2 Cardinality in logging databases

14.2 Lowering the cost of cardinality

14.2.1 Use logging standards to contain cardinality

14.2.2 Using storage-side methods to tame cardinality

14.2.3 Make cardinality someone else’s problem