How Lens Saved Our Time and Created Productivity
At Trendyol, development teams have more than one Kubernetes cluster and several environments (stage, prod, etc.) they have to manage. When the time is to manage or scale the services, we should open an Excel file that includes Kubernetes credentials for opening the Dashboard. In the beginning, it was not a problem and seemed like a normal process. But, managing several clusters at the same time was harder. Your Dashboard access token or session may have expired. You may have copied the wrong token.
Actually, these problems seem not to be a real problem. Indeed, we are human and we can make mistakes simply. At this point, we should make this work simple. Finally, we have found the Lens and we say “That’s it. This is how this work should be done”. Because it is so pragmatic. You can do most of the things (depending on your role) that you have imagined easily without visiting anywhere.
Before getting started, let’s look at how the installation is done.
Installation
Lens is an Electron app. This means any OS that can run Chromium can also run the Lens app. It’s open-source and already has 6.8k stars.
OSX
There is a cask for the Lens on homebrew. You can easily install it with the command below.
brew cask install lensOther operating systems users can download this app from its releases page: https://github.com/lensapp/lens/releases
We are installed successfully. Let’s add a Kubernetes Cluster to the Lens.
Adding a Cluster
Open Lens and choose the + button on the left navigator. As the case may be, the easiest way is to copy your custom kubeconfig and paste it to Lens’s input. We use custom kubeconfig files for using Lens for our clusters.
And, you are ready to go! Let’s look closer.
About Kubeconfig Files
Like other teams, as the OMS team, we are also responsible to manage our applications running on Kubernetes. But, the Kubernetes clusters are shared with other teams. At this point, our developer kubeconfigs should just be able to manage apps (Deployment, Service, ConfigMap, etc.), not the cluster. It’s important for security.
We wanted developer kubeconfigs for seven different clusters from our DevOps team and they gave us.
Moments from Lens
I will show you the moments from my minikube. But, don’t forget there is always more than one cluster in real life. We will address the things we use mostly.
Cluster Overview
At the Cluster Overview, we are looking mostly at summaries of the cluster metrics. We can see CPU/Memory usages, Request, Limits, Capacities, and the Pods.
Also, we can see Cluster level Warnings in the table below.
Node Overview
In this view, we can see the summary of nodes on a list. When you select a node, you can see a view that is opened from the right side covering all information about the selected node.
Workloads Overview
Workloads views are the most used things in our use cases. At the overview, you can see a summary of the workloads, and you can choose one or more namespaces from the right and top sides of the view. Also, you can see an Events table below for selected namespaces. Let’s look at other workloads.
Deployment
We use Deployment objects for most things. On the Deployments view, you can see a table that contains deployments with their summary data. When you click on one of the views that are opened from the right side, it contains all of the information about the selected deployment (ReplicaSets — aka. Revisions, Pods) and you can get the object detail view by clicking.
Let’s scale deployment and see how it is easy on Lens.
At the deployments table, every row has an options button like three points. Click that and choose the scale. Set a scale value and accept it. And that’s it.
There are more workload views, but we will not cover them.
Configuration
In this section, we can edit all the config things like ConfigMap, Secret, Resource Quotas, HPA by selecting one, editing, and clicking the save button.
Conclusion
We have discussed how Lens helped us to manage multiple clusters on one channel and without messing with authentication. This is the most important point for us, and this is why we fundamentally recommend it to you.
If you want to work with large-scale Kubernetes clusters, #cometotrendyol.
