Latency metrics questions

[Lajos]

Hi all,

I was trying to understand how the latency metrics are reported. I 
understand what each of them mean, but do they represent a random 
sampling of timings or some sort of averaging?

If I send through 2000 messages and I see a process_latency of 
1.411764705882353, is that averaged over the 2000 or just for the latest 
message?

As a follow-up, is there a simple algorithm to create the same latency 
figures as are displayed in the UI but using the metrics figures 
themselves? I've tried various things and I can get some figures to line 
up, and others are way off.

Thanks,

Lajos