Skip to content

Glossary

Quick Lookup

Common terms used in this module:

  • Rolling Window - a moving subset of the most recent observations used to compute statistics
  • Rolling Mean - the average of values within a rolling window
  • Window Size - the number of observations included in the rolling calculation
  • Signal - a derived metric created to better understand system behavior
  • Rolling Signal - a signal computed across recent observations to smooth noise
  • Monitoring - observing system behavior over time using metrics and signals

Rolling Window

A rolling window calculates statistics using the most recent N observations. The window "moves" forward one row at a time. Example (window size = 3):

row 1 → mean of [1]
row 2 → mean of [1,2]
row 3 → mean of [1,2,3]
row 4 → mean of [2,3,4]

Window Size

The number of observations used in the rolling calculation. Small windows respond quickly to changes. Large windows smooth the signal but respond more slowly.

Monitoring Signal

A metric used to observe system behavior. Examples:

  • requests per minute
  • error rate
  • average latency

Rolling Monitoring

The process of computing rolling statistics to track how system behavior evolves across observations. Rolling monitoring is a core technique in continuous intelligence systems.