Kubernetes Support about inspectit HOT 2 CLOSED

Thanks for that feature request, this definitely sounds interesting! Do you know how CoScale and Datadog represent this information?

from inspectit.

@pbouillet sorry for taking so long to answer, I was travelling :)

Basically I've seen that CoScale gathers those informations from Kubernetes API and assembles them like this: http://container-solutions.com/content/uploads/2016/09/app.coscale.com-app-00dc3396-54b9-4a9e-9752-13f6c73a872a-dashboard-view-8860-1.png

But this is a general overview, having a view per Namespace and Apps / ReplicaSet is something like this: http://www.coscale.com/hubfs/Blog_Pictures/2016_05/Kubernetes_monitoring.png

You have a nice feature description here: http://www.coscale.com/blog/how-to-monitor-your-kubernetes-cluster

But basically, the component in Kubernetes that gather those informations is Heapster (the CPU/Memory/Disk/Network usage per POD), and you also have the infos about quantity of PODs per RS, PODs running, PODs with problems and also the events directly in Kubernetes API.

You can also take a look in how Datadog works here: https://www.datadoghq.com/blog/monitoring-kubernetes-with-datadog/ and here: http://docs.datadoghq.com/integrations/kubernetes/

Basically it gathers the metric informations directly from cAdvisor running in each Kubelet instance (port 10255)

I think we can start (I can help with some stuffs, but I'm not a pretty good Go developer ;P ) mapping these RS / PODs status, and then we can try later to gather the infos from Heapster also to correlate them all.

What do you think about that?

Thanks!!

from inspectit.