[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Exposing metrics to Prometheus in OKD 3.11

Hi Pavel,

I think that the recommendation with 3.11 is to use the Prometheus Operator [1]. By correctly labelling the service Prometheus is able to discover and monitor the endpoints exposed by all your pods backing the service. More information in the original post [2].



[1] https://github.com/coreos/prometheus-operator
[2] https://coreos.com/blog/the-prometheus-operator.html

On Tue, Dec 18, 2018 at 8:01 PM Pavel Sapozhnikov <pavel sapozhnikov gmail com> wrote:

We have application as follows where every one of our pods exposes a Prometheus http end point to scrape. We had to build our own docker image of Prometheus with static tagerts pointed to service end points. This has worked ok more or less with one exception where if you increase instance of a pod to 2 or more you only still see metrics for one pod because of load balancing. So the solution there is to get it to scrape not services but individual pods. We don't know how to do that, because pods could restart and number of pods can increase per service.

What is the recommended, best approach to get Prometheus to scrape pods? What tools does OKD 3.11 provide with built in Prometheus and Grafana to help us do this, if any?
users mailing list
users lists openshift redhat com

Frédéric Giloux
Principal App Dev Consultant
Red Hat Germany
Red Hat GmbH, http://www.de.redhat.com/ Sitz: Grasbrunn,
Handelsregister: Amtsgericht München, HRB 153243
Geschäftsführer: Paul Argiry, Charles Cachera, Michael Cunningham, Michael O'Neill

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]