Lesson 1: Deciding in advance how each part fails
37 min read·
Module 11 · Lesson 1 · About 50 minutes · Needs: Module 7 lesson 2 (fallback chain, degraded mode), Module 2 lessons 2 and 3 (error prices, residency), the Module 1 failure-mode table. No key and no model calls
The situation
Week 11. Funmi replays 22 September: from 13:00 to 13:40 UTC the primary provider slowed and 30% of its calls hit the gateway's 8-second timeout. Gareth asks what happens when each part fails, not only the model. Leila notes that the first version sent UK questions to the secondary. Helen wants every failure decided in advance.
The running example is trace ask-0646: at 13:12:27 UTC a Tallis Retail Group employee asks "How many days of annual leave do I get?" (the question is attached in this lesson). The primary timed out at 8,000 ms; the secondary answered in 1,929 ms.