I would like to monitor resources(CPU,Memory) on our kubernetes cluster per namespace and container, do you plan do add it directly into stackdriver or hiw can I do it without too much hassle? Thanks
I tried grouping metrics in stackdriver but its missing
look at kube resource explorer. it can list cpu and memory usage at namespace level. follow the link -->https://github.com/dpetzold/kube-resource-explorer
follow the below steps
master $ go get github.com/dpetzold/kube-resource-explorer/cmd/kube-resource-explorer
master $ /opt/go/bin/kube-resource-explorer -namespace kube-system -reverse -sort MemReq
Namespace Name CpuReq CpuReq% CpuLimit CpuLimit% MemReq MemReq% MemLimit MemLimit%
--------- ---- ------ ------- -------- --------- ------ ------- -------- ---------
kube-system kube-scheduler-master/kube-scheduler 100m 2% 0m 0% 0Mi 0% 0Mi 0%
kube-system weave-net-4jb2j/weave 10m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system etcd-master/etcd 0m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system kube-apiserver-master/kube-apiserver 250m 6% 0m 0% 0Mi 0% 0Mi 0%
kube-system kube-controller-manager-master/kube-controller-manager 200m 5% 0m 0% 0Mi 0% 0Mi 0%
kube-system kube-proxy-7275r/kube-proxy 0m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system weave-net-4jb2j/weave-npc 10m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system kube-proxy-jklzm/kube-proxy 0m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system weave-net-s8zd8/weave 10m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system weave-net-s8zd8/weave-npc 10m 0% 0m 0% 0Mi 0% 0Mi 0%
kube-system coredns-78fcdf6894-fg9mv/coredns 100m 2% 0m 0% 70Mi 3% 170Mi 8%
kube-system coredns-78fcdf6894-mw6xc/coredns 100m 2% 0m 0% 70Mi 3% 170Mi 8%
--------- ---- ------ ------- -------- --------- ------ ------- -------- ---------
Total 790m/8000m 9% 0m/8000m 0% 140Mi/17515Mi 0% 340Mi/17515Mi 1%
master $
You can use also prometheus you can find examples here and here
As per documentation
Prometheus is an optional monitoring tool often used with Kubernetes. If you configure Stackdriver Kubernetes Engine Monitoring with Prometheus support, then services that expose metrics in the Prometheus data model can be exported from the cluster and made visible as external metrics in Stackdriver