have you setup readiness and liveness probes for your deployment, if not
please do it. it will send pod into failed state, then your queries will
work.
On Wednesday, August 13, 2025 at 11:29:00 PM UTC+5:30 Akash Kumar Rai wrote:
> Hi,
> We are facing issue with PromQL queries.
> we are using below queries to get the count of total number of pods,
> running pods and failed pods.
>
> total number of pods
> count(max
> without(instance,kubernetes_io_hostname,job)(sum(kube_pod_status_phase{namespace=~"$ns",phase!~"Succeeded"})
>
> by (phase,pod,namespace) > 0)) or vector(0)
>
> running pods
> count(max
> without(instance,kubernetes_io_hostname,job)(sum(kube_pod_status_phase{namespace=~"$ns",phase="Running"})
>
> by (phase,pod,namespace) > 0)) or vector(0)
>
> failed pods
> count(max
> without(instance,kubernetes_io_hostname,job)(sum(kube_pod_status_phase{namespace=~"$ns",phase="Failed"})
>
> by (phase,pod,namespace) > 0)) or vector(0)
>
> however there is a issue with these queries
> even if there are some containers down this query will not count it in
> failed pod
>
> example
> platform-app-86b88c875c-ps99x 0/2 Running 24h
> here the pod status is running but there are no containers up.
> still it is not counted in failed pod
>
> how can i modify the above queries to count total number of pods, total
> running pods and failed pod based on container staus.
>
> Thanking you,
> Akash
>
--
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-users+unsubscr...@googlegroups.com.
To view this discussion visit
https://groups.google.com/d/msgid/prometheus-users/8f1711cf-169d-43f0-afa0-17ebe424d21an%40googlegroups.com.
