Originally created by: mproch
Currently we have "counts" metrics. Maybe we can track also some basic node-latency metrics? The problem is, it can have significant impact on performance. Histograms are not so fast as counters...
Log in to post a comment.
