Kubernetes provides detailed insights about resource usage in the cluster. This is enabled by using Heapster, cAdvisor, InfluxDB and Grafana.

Heapster is installed as a cluster-wide pod. It gathers monitoring and events data for all pods on each node by talking to the Kubelet. Kubelet itself fetches this data from cAdvisor. This data is persisted in InfluxDB and then visualized
using Grafana.

kubernetes-logging

Resource Usage Monitoring provide more details about monitoring resources in Kubernetes. Heapster, InfluxDB and Grafana are Kubernetes addons. They
are enabled by default if you are running the cluster on Amazon Web Services or Google Cloud. But need to be explicitly enabled if the cluster is started using minikube or
kops addons.

Start a Kubernetes cluster on Amazon Web Services as:
KUBERNETES_PROVIDER=aws; kube-up.sh

More details about starting a Kubernetes cluster are available at Getting Started with Kubernetes 1.4. By default, it creates a 4-node Kubernetes cluster in us-west-2a region. More details about the cluster can be seen using the command kubectl cluster-info and it shows the results as:

Note the URL for the Grafana service. Open this URL in a browser window. You’ll be prompted for an invalid certificate warning but this can be safely ignored at this time. In production system, appropriate certificates should be installed.
Then you’ll be prompted for credentials. These can be obtained using kubectl config view command. It will show the output as:

Use the value from username and password fields. This shows the default dashboard:
kubernetes-grafana-empty-dashboard

It consists of two dashboards – one for cluster and another for pods.

kubernetes-grafana-dashboards

For this blog, a 4-node Couchbase cluster was created following the steps outlined in Create a Couchbase Cluster using Kubernetes. A cluster-wide
dashboard shows CPU, Memory, Filesystem and Network usage across all the hosts and looks like:

kubernetes-grafana-cluster

CPU, memory, filesystem and network usage for all nodes may be seen:

Details for each node may be seen by selecting the node:

kubernetes-grafana-cluster-nodelist

CPU, memory, filesystem and network usage for each node is displayed:

kubernetes-grafana-cluster-one-node

Pods dashboard shows CPU, memory, filesystem and network usage for each pod:

kubernetes-grafana-pods

A different pod may be chosen:

kubernetes-grafana-pods-list

A complete list of all services running in the Kubernetes can be seen using kubectl get services --all-namespaces command. It shows the output as:

A complete list of all the pods running in the Kubernetes cluster can be seen using kubectl get pods --all-namespaces.

Some references:

Posted by Arun Gupta, VP, Developer Advocacy, Couchbase

Arun Gupta is the vice president of developer advocacy at Couchbase. He has built and led developer communities for 10+ years at Sun, Oracle, and Red Hat. He has deep expertise in leading cross-functional teams to develop and execute strategy, planning and execution of content, marketing campaigns, and programs. Prior to that he led engineering teams at Sun and is a founding member of the Java EE team. Gupta has authored more than 2,000 blog posts on technology. He has extensive speaking experience in more than 40 countries on myriad topics and is a JavaOne Rock Star for three years in a row. Gupta also founded the Devoxx4Kids chapter in the US and continues to promote technology education among children. An author of several books on technology, an avid runner, a globe trotter, a Java Champion, a JUG leader, NetBeans Dream Team member, and a Docker Captain, he is easily accessible at @arungupta.

Leave a reply