Distributed Systems
Retry amplification: free-recall review
Retrieval beats re-reading. For each prompt, say or write a full answer from memory before you open the model answer — the effort of recall is what makes the mechanism stick when you are paged at 3 a.m.
Reconstruct the unit’s spine — fan-out math, the metastable loop, jitter, retry budgets, circuit breakers, idempotency, and deadline propagation — without looking back at the lesson.
- 01Why do retries compose by multiplication rather than addition, and what is the amplification formula?
- 02Define a metastable failure and explain why the system stays down after the trigger clears.
- 03Why is exponential backoff necessary but not sufficient, and what does jitter add on top?
- 04What is a retry budget, what value is canonical, and what does it bound that backoff cannot?
- 05How does a circuit breaker break the sustaining loop, and what are its three states?
- 06State the two non-negotiable retry rules beyond timing and volume, and why each prevents wasted work.
If you could reconstruct each answer from memory, you hold the unit’s spine: retries multiply (retries^depth), the storm becomes a metastable failure that outlives its trigger, jitter de-synchronizes the herd in time, a ~10% retry budget caps the volume, a circuit breaker grants a zero-retry recovery window through its three states, and the two hard rules — retry only idempotent/retryable work, and propagate the deadline — stop the purest wasted, amplifying calls. The goal is never zero retries; it is retries that cannot multiply into the storm that takes you down.