This will give us time spent in each subsystem per request, on average.
To avoid skewing data for publish request, which normally take much longer time, it is measured separately.
This is version 0.1, if this proves useful I will re-implement metrics collection with db hooks and factor out the systems that need measurement (for example, json marshaling/unmarshalling).
This will give us time spent in each subsystem per request, on average.
To avoid skewing data for
publishrequest, which normally take much longer time, it is measured separately.This is version 0.1, if this proves useful I will re-implement metrics collection with db hooks and factor out the systems that need measurement (for example, json marshaling/unmarshalling).