Prometheus monitoring drilled down metric

Question

Since prometheus is a pulled based metric tracker.

Now lets say I have a simple service and I Need to expose metrics for Prometheus to scrape, Now I do that with one of the clients of the language, configured Prometheus to poll at a frequency of 15secs this API of metrics, Now I have a count metric, Prometheus server at time instance at time t saw count x, At time t+15 sees count y

Now becuase of the scrap interval(pull model) I don't have any idea wether the increase of count x to y had a steady slope or a Spike of requests came in at t+10

Answer 1

在prometheus组获得了答案https://groups.google.com/forum/#!topic/prometheus-users/8--nJRokE8Y