get CPU usage of all the namespaces

[Darshil Saraiya]

I am using prometheus 1.7.0 and kubernetes 1.6.7.

I used memory metrics and got results of all the namespaces

sum by (namespace) (kube_pod_container_resource_limits_memory_bytes)

Element	Value
{namespace="kube-system"}	178257920
{namespace="monitoring"}	1029701632

So, I got different namesapces in this query

But, when I used cpu metrics, I got the results only containing "monitoring" namespace:

sum by (namespace) (kube_pod_container_resource_limits_cpu_cores)

Element	Value
{namespace="monitoring"}	1.525

 

Where I have these namespaces:

$ kubectl get ns

NAME          STATUS    AGE

default       Active    1h

kube-public   Active    1h

kube-system   Active    1h

monitoring    Active    25m

So, Should I use this https://www.weave.works/blog/aggregating-pod-resource-cpu-memory-usage-arbitrary-labels-prometheus/ and add Scarping in config file and then get the info about different namespaces or is something not right in these results?

[Callum Styan]

Please use prometheus-users (https://groups.google.com/forum/#!forum/prometheus-users) for usage related questions. This mailing list is for discussions related to development of Prometheus itself.

--
You received this message because you are subscribed to the Google Groups "Prometheus Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to prometheus-developers+unsub...@googlegroups.com.
To post to this group, send email to prometheus-developers@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/prometheus-developers/98877768-fb8d-449e-ae67-fd337bbc3a7d%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.