How would you implement exponential backoff with randomized jitter?

Exponential backoff with jitter is a retry strategy that increases the wait time between attempts exponentially while adding randomness to avoid synchronized retries. The algorithm:

1. Set a base delay (e.g., 200 ms) and a maximum delay (e.g., 30 s).
2. For each retry attempt i (starting at 1), compute a delay = min(maxDelay, baseDelay * 2^(i-1)).
3. Apply jitter by multiplying the delay by a random factor between 0.5 and 1.5 (or using full jitter: random(0, delay)).
4. Sleep for the jittered delay before retrying.

This approach reduces the probability that many clients hit the service simultaneously, mitigating the thundering herd problem. In AWS SDKs, you can configure the retry strategy by providing a custom retryer that implements this logic. The time complexity is O(1) per retry, and the expected number of retries is bounded by the maximum delay. This pattern is widely used for transient errors such as throttling, network glitches, or temporary service outages.