Exponential Backoff

A retry timing strategy where the wait interval doubles after each consecutive failure — for example, 1 s, 2 s, 4 s, 8 s — capping at a configured maximum. This prevents a flood of retries from overwhelming an already struggling server. Exponential backoff is almost always combined with jitter to desynchronize retries from multiple clients. It is the standard approach recommended by AWS, Google, and most distributed systems literature for handling transient errors.