AI Engineering · Request-Path Reliability

Reliability engineering for a single LLM call, from sending the request to receiving the result: choosing a model, falling back after failure, and keeping timeouts and retries from amplifying an outage.

Chapters

1. Chapter 3: Model Gateways, Routing, and Fallback
2. Chapter 4: Retries, Timeouts, Idempotency, and Circuit Breakers

Back to AI Engineering / LLMOps.