Understanding Software Dynamics

Chapter 13

Observation Tool Design Principles

In the first part of this book, we learned how to do careful measurements of four of the five fundamental shared computer resources: CPU, memory, disk/SSD, and network (we defer discussion of critical section locks to Chapter 27). We learned about displaying measurement results in useful ways. We observed variations in transaction latency in Chapter 7, and unexpected software dynamics in Chapters 5 and 7. We touched on transaction delays on one machine stemming from waiting on RPCs to other machines, from interference by other programs and the operating system on the one machine, and from unreasonable offered loads.

In this second part, we have looked at tools for observing software dynamics—logs, ...