KUBEVIEW KUBERNETES
Visualize your Kubernetes Real-Time
Monitor your cluster visually with KubeView
Kubernetes or k8s solves one of the biggest problems that ever existed- Scaling. The auto horizontal scaling of its pods with the best utilization of given resources makes it utilize the whole resources very efficiently. Based on the concept of dockerization k8s makes the whole app host platform-agnostic thereby building trust to run your software correctly anywhere and not just ‘on my machine’.
Though having a steep learning curve, the result is quite interesting and peaceful. You now don’t have to worry about going down of your server or bare metal as k8s manages that on its own.
Problem
One of the problems we faced just after finishing the k8s setup is to monitor the cluster. Prometheus and Grafana seem to provide a pretty good monitoring dashboard but we looking more into where we can get a full visual overview of what's happening inside the cluster. We can summarize the requirements better:
- Hows many services, deployments are there.
- Visual of our ingress controller.
- The number of pods in each service and their health.
- Architecture diagram of our real-time cluster.
KubeView
Upon exploring new tools, we came to KubeView. This exactly provides what we needed. A real-time visual diagram of different k8s objects in our cluster.
Here, we can easily see different pods, services, deployments, ingress, IP, lets-encrypt clearly.
Healthy ones are indicated by green color while unhealthy ones are shown in red and waiting/creating ones in grey.
Also, we get a clean architecture diagram of our cluster with real-time stats.
Filters for the namespace, refresh intervals are provided for the monitoring needs.
Deploy
KubeView can be easily deployed using its helm chart. There are two ways to do this:
- By directly using helm chart:
$ helm repo add cowboysysop https://cowboysysop.github.io/charts/
$ helm install my-release cowboysysop/kubeview- By using the helm chart provided in the repository (latest version)
# Clone git repo
$ git clone https://github.com/benc-uk/kubeview# Navigate to kubeview/charts
$ cd kubeview/charts# Deploy the chart
$ helm install kubeview-release ./kubeview
You will see Kubeview service will be up and running. And to view the board you need to select KubeView pod and port forward it. Visit localhost:8000 to view your KubeView dashboard.
To expose your KubeView service to an external IP you can set custom parameters while installing your helm chart:
$ helm install kubeview-release ./kubeview --set service.type=LoadBalancerThe LoadBalancer service type will expose your KubeView service to an external IP.
It's always better to deploy your chart in a separate namespace and for that, you can pass the following parameter:
$ helm install kubeview-release ./kubeview --set service.type=LoadBalancer --create-namespace -n kubeviewThis will now create a namespace named kubeview (if it doesn’t exist) and deploy the chart in that. Make sure you are on this namespace to view its pods.
Conclusion
Though there are many different powerful ways to monitor your cluster, for us KubeView seems the best to get insights in one to go so as to what happens in real-time.
Github Repository: https://github.com/benc-uk/kubeview
Whola! Both you and I learnt something new today. Congrats
Clap! Clap! Clap!
Connect with me:
Portfolio: https://apgapg.github.io/
LinkedIn: https://www.linkedin.com/in/ayushpgupta/
Github: https://github.com/apgapg
