Sampling
In plain terms
Recording a representative subset of data instead of all.
Going deeper
Sampling captures only a fraction of events (like traces) to reduce overhead and cost while preserving useful insight. It trades completeness for efficiency.
Under the hood
Head vs. tail sampling decide which traces to keep; balances visibility against volume.
Examples
Distributed tracing, high-volume logging, cost control, statistical analysis.
Think of it like...
Like surveying a representative slice of a crowd instead of every single person.