I am developing a large scale distributed system written in perl with multiple machines performing different data processing tasks. Modules crash from time to time and when they do, I write out their stacktrace to a log file. I later scan the log for crash patterns.
Is there a more reliable way (or even standard practice) of collecting bugs, filing and reporting them? I am looking for perl packages or software that can do the bug collection and processing tasks.
In particular I were to write one myself, what are some of the features a crash collection system should consider? My homegrown bug collector records the buggy input, stacktrace, machine etc. Upon hitting a bug the
module restarts on the next input and resumes processing.
Any advice would be much appreciated.