Stuff Goes Bad:Erlang In Anger


Review Questions

[1]. What are the two main approaches to pin issues about CPU usages?
[2]. Name some of the profiling tools available. What approaches are preferable for production use? Why?
[3]. Why can long scheduling monitors be useful to find CPU or scheduler over-consumption?

[2].列出一些可用的profiling 工具。生产环境中,那一种方式更加适合?为什么?
[3].为什么long scheduling monitors可以有效地找出CPU或调度器过度消费问题?

Open-ended Questions

[1]. If you find that a process doing very little work with reductions ends up being scheduled for long periods of time, what can you guess about it or the code it runs?
[2]. Can you set up a system monitor and then trigger it with regular Erlang code? Can you use it to find out for how long processes seem to be scheduled on average? You may need to manually start random processes from the shell that are more aggressive in their work than those provided by the existing system.