How to measure Errors of a Kubernetes Service with Prometheus

[Ness]

Let's say I want to measure 40x and 50x errors of a K8s service "frontend-service" example, whats the best way of targeting this, not much info on google.

Thanks

[Ian Billett]

Hello Ness,

Services in Kubernetes are kind of like load-balancers - they just route requests to underlying pods. The pods themselves actually contain the application that does the work and returns the status code.

You don't monitor kubernetes services per-se for 4xx or 5xx errors, you need to monitor the underlying application itself. This is why there will be not much info on Google 👍

Hope that makes sense.

Best,

Ian

On Fri, Jul 9, 2021 at 12:04 AM Ness <ne...@cuburn.com> wrote:

Let's say I want to measure 40x and 50x errors of a K8s service "frontend-service" example, whats the best way of targeting this, not much info on google.

Thanks

--
You received this message because you are subscribed to the Google Groups "Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-users/9ee2df6f-7f96-4a8e-8f27-0bf6286ac39en%40googlegroups.com.