Timeouts Trial: Stop One Slow Dependency From Freezing Everything

One dependency gets slow, then the whole application starts behaving badly. Participants diagnose cascading latency and apply practical controls: request timeouts, retry budgets, jitter, circuit breakers, bulkheads, and user-facing fallback states.

The lab keeps the pressure visible with dashboards and request traces, then asks teams to justify each setting. The point is to learn why reliability settings are product decisions, not random numbers copied from a blog post.