Kubernetes in the cloud: Architecture and Overview


With the recent surge in popularity of Kubernetes among IT operations teams and developers, there is a rising demand for Certified Kubernetes Administrators (CKAs) or professionals with CKA certification. But, let us first know about Kubernetes and its various other aspects.

What is Kubernetes?

For developers and IT operators looking to automate manual processes such as management, deployment, scaling, and routing of containerized applications, Kubernetes offers an open-source platform for container orchestration. In simple words, Kubernetes aids in the smooth and efficient management of a cluster of hosts running Linux containers. Kubernetes technology is a product of Google lab that can manage host clusters in virtual, physical and cloud environments. It is perfect for hosting cloud-native applications spanning across private, public or hybrid clouds.

Kubernetes has several capabilities that include – service and process definition, storage orchestration, scaling of containerized applications, automated deployment and updates, container health check, container-level resource management, and others. With the help of Kubernetes, developers can get a constant infrastructure for running distributed systems, right from development to production. However, Kubernetes’ orchestration services are dependent on other open-source projects that support features like registry, networking, telemetry, security, and automation.

Kubernetes architecture

The first step to a CKA certification is getting acquainted with the Kubernetes architecture, which will be discussed in this section.

 A cluster is a working Kubernetes deployment. Besides requiring persistent storage, container registry, and underlying infrastructure, it has the following two parts:

  1. Control plane – The control plane consists of master nodes, where the core Kubernetes components control the cluster states and configurations. The elements of the master node include:
  2. API server – It provides an interface for interacting with the clusters and communicates with different libraries and tools. API can be accessed through REST calls or through the Kubeconfig package that comes with server-side tools.
  3. Etcd – It is a key-value store that holds all the information regarding the state of the clusters.
  4. Controller manager – It is an amalgamation of several controller functions for collecting and transmitting data to the API server. Key controllers include endpoint, replication, namespace, and service account controller.
  5. Scheduler – It assigns pods to available worker nodes.
  6. Worker nodes – Each node could either be a virtual or physical machine and runs pods, which consist of containers. Worker node components are:
  7. Container runtime engine (Dockers) – Each node requires a docker that helps in running the containerized applications in a suitable operating environment.
  8. Pods – It represents a single unit of an application.
  9. Kubelet –  It relays information and communicates with the control plane.
  10. Kube-proxy – It carries out networking services and makes services accessible to external hosts.

Setting up Kubernetes on cloud

Pre-requisites of running Kubernetes on Google Compute Engine:

  1. Google Cloud Platform Account
  2. Installed gcloud (with authentic credentials) that can use your Google Cloud Platform project.
  3. Enabled Compute Engine Instance Group Manager API
  4. A GCE VM from the command line
  5. Ability to SSH into the VM without prompts

Step 1 – Installing a client and starting a cluster using either of the following:

curl -sS https://get.k8s.io | bash


wget -q -o – https://get.k8s.io | bash

The script for cluster startup starts a running Kubernetes cluster and leaves a Kubernetes directory on the workstation.

Step 2: Installing the ‘kubectl’ command-line tools on your workstation using gcloud.

gcloud components install kubectl.

Step 3: Cluster inspection such as ‘kubectl get –all-namespaces services’  for looking at services and ‘kubectl get —all-namespaces pods’ for inspecting pods.

Step 4: For removal, deletion or tearing down of the cluster, the command can be used.  The same directory contains ‘kube-up.sh’ that will bring back the cluster.

Step 5: Default parameters can be customized using the command:


Guide to troubleshoot:

-For project settings, besides the parameters mentioned in the pre-requisites, you have to enable the Google Cloud Storage JSON API, and Google Cloud Storage API.

-In the event of a Kubernetes cluster startup script hanging, SSH into the node VMs, and the master node and check logs using /var/log/startupscript.log. 

-After the issue is resolved, run ‘kube-down.sh’ before running ‘kube-up.sh’.

-If the default networking rules are missing, create a new rule using field values:

  • Source Ranges
  • Allowed Protocols and Port top:1-65535;udp:1-65535;icmp

Source: https://kubernetes.io/docs/setup/production-environment/turnkey/gce/

Need to learn Kubernetes and become CKA certified

The growing demand for Kubernetes calls for skilled professionals with CKA certification. The Certified Kubernetes Administrator or CKA program is a joint effort of the Cloud Native Computing Foundation and the Linux Foundation to develop the Kubernetes ecosystem and train individuals in the Kubernetes domain. A thriving community of Kubernetes Administrators will ensure the continued growth and development of companies that use Kubernetes on a large scale. A CKA certification is a testimony to the professionalism and reliability of a Kubernetes Administrator that adds on to their value and demand in the job market. On the other hand, IT operators and developer companies that require Kubernetes can quickly identify skilled and talented administrators.

The certification program provides hands-on training in a command-line environment to ensure that the CKAs have the requisite knowledge, expertise, and proficiency in providing the leadership that an administrator requires. The program is conducted online through performance-based tests that require the users to solve several problems from a command line within three hours. The course is designed to bring out the skills of a candidate that are essential to be a champion as a Kubernetes Administrator in today’s industry. The examination tests the ability of the candidates in some general domains like core concepts, installation, configuration and validation, management of application lifecycle, networking, scheduling, security, and troubleshooting.


With several advantages including high built-in fault tolerance and automation capabilities, Kubernetes has proved to be very useful for companies intending to replace manual tasks with automation.  It can run applications on virtual, physical, as well as a cloud infrastructure that has eased the management and deployment of applications that are containerized. All in all, there cannot be a better time to utilize your knowledge of Kubernetes to make a mark in the industry through a certified CKA program.