Blog

Thinking out loud about backend systems

Notes on architecture decisions, scaling challenges, and lessons learned from building systems in production.

No hot takes. No hype. Just what I've found useful.

Production Engineering (1)

Diagram showing one request growing to 64 calls at a payment gateway when three service layers each retry three times

Retries That Make Outages Worse

Retries are supposed to make systems more reliable. Under load, they're often what keeps an outage going. How retry storms happen, and how to design retries that back off instead of piling on.

retriesresilience+5

8 min readSep 22, 2026

Stay Updated

Get notified when I publish new posts. No spam, unsubscribe anytime.