Totally agree. There’s a lot of literature and hacks on what to break, degrade, etc. Hardly anything on the fact that you have to have good performance and faul

Totally agree. There’s a lot of literature and hacks on what to break, degrade, etc. Hardly anything on the fact that you have to have good performance and fault tolerance requirements - which come with workload and architectural constraints. Also hardly anything on good statistical methods to compare metrics and which metrics to compare