Performance

Jitter

Variance in latency or arrival rate.


In plain terms

A constant 100ms is fine; latency that swings 5–500ms is debilitating. Sources: GC pauses, network microbursts, scheduling delays.

Origin

Term as old as queueing theory. AWS's "Exponential Backoff and Jitter" (Marc Brooker, 2015) is the canonical write-up of why jittered retries beat un-jittered ones every time.

Where it shows up in production
  • AWS SDK All AWS SDKs use jittered exponential backoff by default.
  • Erlang/OTP Built-in supervisor restart intensity has a randomization parameter for the same reason.
Sources & further reading
Found this useful?