Kubernetes: The Captain of the Docker Ship


The name Kubernetes originates from Greek, meaning helmsman or pilot. A Google brainchild, that was open-sourced by Google in 2014, Kubernetes is truly a driving force. It is responsible for the seamless deployment of codes in software development. How is Kubernetes making our lives better? Let’s find out.

You could run a single instance of any application by a simple Docker run command. But that’s just one instant of one Docker host. What happens when multiple hosts try for that instant? Obviously! It cannot entertain the workload anymore. So, you start deploying more instances by individual Docker run command. As simple may it sound, the process altogether can be of a lot of fuss. To begin with, you will have to keep a close tab on the Docker host to keep adding instances whenever required. Also, the health of the instances is of utmost importance. Well said that the docker itself could just out of the blue fail. Hence, making the instances inaccessible. 


Understanding Kubernetes

To understand these concepts, take an example of HOOQ, a digital streaming platform. HOOQ experiences a certain amount of traffic on weekdays, whereas traffic during weekends is almost 5x to 6x. Now, suppose HOOQ servers are equipped enough to handle the weekday traffic. But all the login requests or the sudden surge in traffic during the weekend is too much for its servers. It can suffice once or twice but then that’s it. The servers are likely to crash after a point. This is the point where scaling-up comes to your rescue. Yes! You thought about right Docker can scale-up its containers when the workload is more. But remember that process is manual. Therefore, it would more be like HOOQ putting up extra servers and all to scale-up the workload for the weekends and then destroy the infrastructure to scale-down the containers in the weekdays. That’s downright not feasible! This is where container orchestration walks down the aisle. This a set of scripts, and tools that assist the host containers to survive in a production environment. It keeps track of the traffic and automatically scale-up the containers once the threshold is reached. Also, it contains multiple Docker hosts. Therefore, even if an entire Docker host fails, the instances are still accessible from others.

I assumed all the factors in HOOQ for your understanding, don’t go places to authenticate the data. So, coming back to what Kubernetes is- Kubernetes is an automated, extensible, open-source platform for container orchestration. It is a platform managing containerized workloads and services. Along with that it also balances the workload across all the containers.


Components of Kubernetes

Let me now walk you through the components of Kubernetes. When you first deploy a Kubernetes, what you get is a cluster. This cluster contains nodes, these are worker machines running the containerized applications. These nodes host the components of the application workload, pods. There is a control plane that manages these nodes and pods in a cluster. That panel is known as the master. Simply put, you run your containerized apps in nodes and you control them through the master.

But before you get to know what Kubernetes is, it’s better to talk off about the myths and misconceptions that surround the term. Well, to start with Kubernetes is not a containerization platform, it is a containerization management tool. In a continuous development structure like DevOps, it is very important to retain consistency across all the servers. Therefore, such a tool is extensively important to simultaneously deploy the codes. Another important factor is that it is not comparable to Docker. Remember, Kubernetes works by managing the dockers. Containerizing remains with Docker, or Rocket maybe even Linux containers as per the company’s discretion. Kubernetes is as its name suggests the captain, of Docker ship.


Features of Kubernetes

Next on my list are the features of Kubernetes

  • It is quite a task to place the containers in an orderly fashion. This feature automatically packages the application by scheduling the container according to their requirements or whatever information available on them.
  • It gives you complete freedom to opt for any storage system you want or require. Be it local storage, or a public cloud provider such as GCP or AWS, or perhaps a shared network storage system such as NFS, iSCSI, etc. it can suffice them all.
  • Kubernetes keep close tabs on the health of the nodes and containers. This is the point that explains our concern that what happens if a node or container dies or fails to execute. Well to start with it is capable of restarting containers once they fail to execute. It might even kill the containers that do not stand up to the health standards. And for the dead nodes, it can execute them back on available Docker containers.
  • We know load balancing and cloud directory, i.e. where you store the containers are a huge deal. You can technically store your containers wherever you desire but that makes complicated to find and schedule for deployment. With Kubernetes, you can relax, because networking is taken care of. It only assigns IP addresses for the containers. Also, a single DNS is allocated to a set of containers so, as to distribute the traffic or workload in a required manner.
  • Kubernetes allows you scaling resources not only vertically but also horizontally, easily and quickly.



Before I conclude this article, I would like to refer to another point, we already discussed the fact that Docker is not comparable to Kubernetes! So, what was the hype about in the market? The right comparable application is Docker Swarm. These are popular tools for container management and orchestration. I know the logical cells in your brain are collectively zeroing upon the leverage that Docker Swarm should have upon Docker. But let me assure you Kubernetes is the clear winner. Mostly due to the auto-scaling feature.

As the market is growing, so growing is the ever-evolving requirements. Dockers were amazing, still are but the industry needed that extra something. Also, with the flourishing of continuous software development, the deployment phase is pacing up. To do that the changes to the code should be smooth yet efficient enough to reflect on the user-end. This kind of orchestration not only performance of the application but also slashes down its time to marker.

Looking for some more in-depth grip on Kubernetes? But confused about which course to take-up? Or you are an industry expert looking? Then look no further because Nuvepro provides hands-on labs to help complete your learning. If you are in the process of learning the concepts of Kubernetes or may be thinking of taking it up. Even if you are an expert looking to deploy these tools for your teams, please reach out to us, we will be happy to help.









Kubernetes: The Captain of the Docker Ship


DevOps as it’s said in not only a methodology but a mindset as well. Therefore, several teams need to work as a union of development and operations. Also, DevOps has quite a few components that need to merge seamlessly to get what is required. A key element, which usually serves as the center of the DevOps “structure,” is configuration management.

So, what is this configuration management?

Configuration management can refer to something unique in different industries. In the software industry, it refers to managing any item that is configured for the sake of the project. For example, source codes, property files, binaries, servers, and tools can all be configuration items for a software firm. This solely can decide the future of the project. As we know for software to be developed successfully various factors are taken into account. But the most critical is the configuration as it varies in all aspects.

Comprehensive Configuration Management

DevOps, as we know, is an alliance between development and operations. Therefore, it is only obvious that configuration management also spans over both sections.

Configuration management ranges over three broad sections

  • Artifact Repository- It is a database of files that are not always in use. Like libraries, or test data. DevOps is a continuous process, so, it is always producing such files. These required to be stored but not necessarily be accessed.
  • Source Code Repository- It is the opposite of the above. This is a database to store the working codes. It also stores configuration files and various scripts.
  • Configuration Management Database- It is a relational database that ranges over several systems and applications related to configuration management, including services, servers, applications, and databases, and many more.


Configuration Tools- Puppet, Ansible, and Salt Stack

With the basic know-how of configuration management. Let us look into a few tools used in this process: Puppet, Ansible, and Salt Stack. For that, we will take a comparative approach. We will be looking into these tools concerning a few metrics:

Availability– These tools are the heart of servers of huge enterprises. The entire development process depends on the configuration because from source codes to the prod servers to the hosts all of them are controlled by these. Therefore, it is very important that whenever any part of it fails, back-up is readily available.

  • Puppet- It works on master-slave architecture. It has a multi-master approach. Whenever the active master crashes, the other takes over the charge.
  • Ansible- It has only one active node, primary instance. But it does provide a secondary instance in any such cases for back-up.
  • Salt Stack- It is capable of configuring multiple masters. Hence, it has multiple main servers to maintain the configurations across.

Ease of Setup– Though this is a small price to pay for all the hectic workload that gets automated with the use of these tools. But for someone doing this for the first-time ease of setup does matter.

  • Puppet- It has multiple masters at its call. Also, Puppet’s main server runs on the master machine and Puppet client runs as a slave on each machine. Therefore, the set-up seems quite a task.
  • Ansible- It works on a single active- node, and no agents running on the slave machines. Therefore, no separate VM is required on each of the client machines. This approach makes it easier in terms of setup.
  • Salt Stack- It also works on the master-slave architecture, it also, has multiple servers at its disposal. Host servers are known as salt masters, and clients are called salt minions. It is also a heavy setup to succumb.

Management– There are two types of configurations push and pull. The push configurations principle is that the servers push the configurations to the nodes connected. In pull configurations, the nodes pull the configurations from the central servers to themselves without any command.

  • Puppet- It follows a pull configuration. It has its unique Domain-Specific Language- Puppet DSL. It is known as the system-administration oriented tool. Immediate remote execution is also not available in Puppet.
  • Ansible- It follows the push configuration. It uses YAML- Yet another Mark-up Language, Python which is quite user-friendly. Also, it provides immediate remote execution.
  • Salt Stack- It also follows push configuration, and uses YAML.

Interoperability: Another Aspect is the interoperability of all the above-mentioned tools. All of these tools have the same condition. The main/central server or the master has to operate on Linux/Unix. Whereas the slave or client machines can work on windows.

Scalability: Scalability is another highly valuable aspect. All of these tools are equally well equipped to handle any kind of scalability whether scale-up or scale-down. You would have to just specify the IP addresses and the hostname of the nodes to be configured.


Functions of the Configuration Management Tools

Now after comparing all the tools, we have understood these are responsible for deploying, configuring, and managing the servers. Now, I will list out the functions these tools perform:

  • They provide centralized control over all the machines configured through any of these tools. Hence, no fuss in implementing any change. Any change in the main server can propagate the change in all.
  • Defines the configuration of the host servers, also keeps close tabs for any changes. In case of any changes, they are equipped to fall back to the default configuration.
  • It can scale-up or scale-down servers depending on the requirement.



DevOps starts with configuration and ends with it too. Also, in between the start and end, it remains configured. The main objective of DevOps is to develop software at the earliest. This calls for an outstanding and flawless (well almost) approach. Configuration management is an indispensable part of that approach. We cannot forget that even if we can replicate the client environment by placing our applications in containers but that is not enough. The configurations of each machine will always differ. We cannot separate configuration management from DevOps, doing so would be a faulty perspective towards DevOps.

All buckled up to explore this awesome feature? Then look no further because Nuvepro provides hands-on labs to help complete your learning. We can help you master the concepts of Configuration Management. Already an expert? Well, we have got you covered too. Deploy these tools to your teams with Nuvepro for a seamless experience. Please reach out for further details.

Key references:




For Enterprises

Upskill and reskill your employees and make them ready for new projects.

For Online Learning Providers

Enhance the learning experience and learner stickiness

For Educational Institutions

Enable students to rapidly get skilled and be industry ready.

See a real-life use case that is relevant to your organization