We've been running 16.01 for more than a month now, and our previous version was 15.07. As 16.01 use Nginx we have logs with request time and we noticed that there is a big difference between the times that those logs are reporting and what we see in the Analytics Database, sometimes are seconds. We're using this to calculate the elapsed time on the db.
A few questions related to that.
1) What's the scope of the Nginx logs? Is it first byte of the coming payload + MP + target and first byte delivered to the client?
2) Which times is reporting the Analytics Database? Is it the first byte to the MP + target and and first byte delivered to the client?
We're using Graphite / Grafana integration to monitor activity on Apigee and since the upgrade we noticed that the percentile 95 of those graph changed. On 16.01 it shows that the times are faster but as @Eric Dahl asked before here we're not really sure what that metric is reporting. Is it possible that previous to the use of Nginx the time of the router was included in the Analytics Database and now with Nginx it's not. Also, are the metrics from Eric post calculated based on which components? The fact that they seem similar to what we have in the analytics db makes me think that those are from MP + target and back.
This would help us to understand what the scope of those metrics are.
LinkedIn
Twitter
