Prometheus metrics are not matching with kubestate metrics in kubernetes dashboard

0
votes

Metrics in Kubernetes Dashboard: enter image description here

kubectl top report:enter image description here

Metrics in Prometheus Grafana Dashboard: enter image description here

I installed Prometheus operator setup using helm chart prometheus-community/kube-prometheus-stack. Can anyone explain me why there a difference in the metrics ? As both prometheus & kubernetes dashboard use kube-state-metrics. which one is better to use ?

1
kubernetesgrafanaprometheus-node-exporterkube-state-metrics
What exact command do you use, what kind of query did you apply using prometheus ? Please paste commands and arguments you have used. – Malgorzata
Memory Usage Query in Grafana : sum(container_memory_working_set_bytes{cluster="$cluster", namespace="$namespace", container!="", image!=""}) by (pod) – Santhosh reddy
Did you take a look github.com/lensapp/lens/issues/607 ? – Malgorzata

1 Answers

0
votes

I think that on the K8s dashboard you are looking at aggregated for 7 days, and on the grafana for the last value of the metric? But it can be something else there is too much info missing. Are we talking about a pod or node? does kubectl top report something totally different?