Kubernetes Interview Questions

⌚ What do you mean by Kubernetes?

An open-source container management platform called Kubernetes is in charge of load balancing, scaling and descaling of containers, and container deployment. Being a Google creation, it provides excellent community and functions flawlessly with all cloud providers. We can therefore conclude that Kubernetes is a multi-container management tool rather than a platform for containerization.

⌚ Describe a container?

A mechanism for storing compiled code for an application until it is needed during runtime is known as a container. Every container enables you to execute recurrent, standardised dependencies and consistent behaviour. It separates the application from the supporting host infrastructure to make cloud or OS platform deployment more simpler.

⌚ How are Docker and Kubernetes related?

It is common knowledge that a Docker image creates runtime containers and that Docker manages the lifespan of containers. However, Kubernetes is employed because these distinct containers need to communicate. Therefore, containers are created by Docker, and they communicate with one another using Kubernetes. Therefore, Kubernetes may be used to manually link and coordinate containers that are running on many hosts.

⌚ Explain the difference between deploying applications on hosts and containers?

Applications that are deployed are made up of an architecture and an operating system. The operating system's kernel will contain the numerous libraries that are installed on the system and are required by an application. The machine that runs the containerized processes is referred to as a container host. The apps must have the required libraries because this kind is segregated from other applications. The binaries can't violate any other application because they are isolated from the rest of the system.

⌚ What characteristics does Kubernetes have?

The user has control over which server will host the container thanks to Kubernetes. It will determine the launch method. Kubernetes thereby automates a number of manual operations.
Kubernetes simultaneously maintains several clusters.
It also offers a number of extra services, including networking, storage, security, and container management.
The health of nodes and containers is self-monitored by Kubernetes.
Users may rapidly and easily grow resources horizontally in addition to vertically with Kubernetes.

⌚ What are the components of Kubernetes architecture?

The master node and the worker node are the two main parts of the Kubernetes architecture. There are separate components in each of these parts.

⌚ How is containerized deployment made simpler by Kubernetes?

All of these containers would need to communicate with one another because a typical application would have a cluster of containers running across several servers. So, in order to do this, you need a large-scale system that can scale, load balance, and monitor the containers. Your choice must be Kubernetes, which can run on any public or private provider and is cloud-agnostic, to make containerized deployment simpler.

⌚ What do you mean by node in Kubernetes?

The smallest basic piece of computing hardware is called a node. It depicts a single computer in a cluster, which could be either a physical computer in a data centre or a virtual computer from a cloud service provider. In a Kubernetes cluster, each machine can replace any other machine. The Kubernetes master manages the nodes that house the containers.

⌚ What are the functions and services that a node provides?

The following services are provided by a node:

Container run-time
Kubelet
Kube-proxy

Starting and managing containers is the responsibility of the container run-time. Each node's state is managed by the kubelet, which also manages the pods' metrics collection and takes commands from the master to operate on each node. A component that oversees subnets and makes services accessible to all other components is the Kube-proxy.

⌚ What activity occurs on the Kubernetes Master Node?

On the master node, the Kube-api server process is running and helps scale the deployment of more instances.

⌚What exactly are pods used for?

A group of containers that are deployed on the same host as a pod are known as pods. The Kubernetes unit of object models can be created or deployed by the Kubernetes application's fundamental execution unit.
There are two ways to use Kubernetes pods. They are listed below:
runnable pods that fit into a single container
Pods that can operate alongside many containers when cooperation is necessary

⌚ What are the orchestration tools for containers?

A framework for managing microservices and containers at scale is offered by various container orchestration technologies. The following are the most often used tools for container orchestration

Kubernetes
Docker swarm
Apache Mesos

⌚ Could you give a brief explanation of how Kubernetes' master node functions?

The nodes are under the direction of the Kubernetes master, and the nodes contain the containers. Now, these individual containers are housed inside pods, and depending on the configuration and needs, there might be a different number of containers inside each pod. The pods can therefore be deployed using either a command-line interface or a user interface, depending on the situation. Then, these pods are scheduled on the nodes, and the pods are distributed to these nodes according on the resource needs. The kube-apiserver ensures that communication is established between the master components and the Kubernetes node.

⌚Describe ETCD?

Etcd is a distributed key-value store for managing distributed work that is written in the Go programming language. Etcd, which represents the status of the Kubernetes cluster at any given time, holds the configuration data for the cluster.

⌚ What is the Kubernetes controller manager?

The Kubernetes Controller Manager is a single process that runs as a compilation of the several controller processes running on the master node. As a result, Controller Manager is a daemon that creates namespaces, manages trash collection, and embeds controllers. In order to manage the endpoints, it is in charge of doing so and interacts with the API server.

⌚ What do you mean by Heapster?

Heapster is a node-based cluster-wide data aggregator that uses Kubelet to collect data from each node. The Kubernetes cluster natively supports this container management tool, and it functions as a pod exactly like any other pod in the cluster. In essence, it finds every node in the cluster and uses an on-machine Kubernetes agent to query usage data from the Kubernetes nodes in the cluster.

⌚ Describe Kubelet?

The slave can communicate with the master thanks to an agent service that runs on every node. In order to ensure that the containers described in the PodSpec are healthy and operational, Kubelet works with the description of the containers that was provided to it in the PodSpec.

⌚ How does the Ingress network operate and what is it?

Users can access your Kubernetes services from outside the Kubernetes cluster thanks to an item called an ingress. By establishing rules that specify which inbound connections reach certain services, users can configure access.
How it functions: This API object provides the routing rules to control how external users access the services in the Kubernetes cluster via HTTPS or HTTP. With this, users can build up the traffic routing rules quickly and efficiently without having to make numerous load balancers or expose every service to the nodes.

⌚ Mention the Kubernetes object list?

The list of objects that are used to specify the workloads is provided below.

Pods
Replication sets and controllers
Deployments
Distinctive identities
Stateful sets
Daemon sets
Jobs and cron jobs

⌚ What distinguishes the pod from the container?

Pods are the grouping of containers utilised by Kubernetes as the replication unit. Containers are the group of codes that must be compiled in an application's pod. Containers in the same pod can communicate with one another.

⌚ What does Kubernetes' stateful sets mean?

A workload API item called a stateful set is used to manage stateful applications. It is employed to scale the sets of pods and manage deployments. In the disc storage that connects to the stateful set, the state information and other resilient data of stateful pods were saved and kept up to date.

⌚ Describe the Kubernetes architecture?

Master Node: The initial and most important part of the Kubernetes cluster, the master node is in charge of managing the cluster. It serves as the starting point for all administrative tasks. To test for fault tolerance, the cluster may have more than one master node.

API Server: The API server serves as the starting point for any REST commands required to manage the cluster.

Scheduler: Tasks are scheduled to the slave node by the scheduler. Every slave node's resource utilisation data is kept there. It is in charge of allocating the task.

ETCD: wright values, store configuration information, and etcd components. For work and to receive commands, it communicates with the majority of the component. Additionally, it controls port forwarding operations and network regulations.

Worker/Slave nodes: These nodes are still another crucial part that manages the networking between containers, interacts with the master node, and enables you to provide resources to the scheduled containers.

The Kubelet: It obtains a Pod's configuration from the API server and makes sure the mentioned containers are operational.
Docker Container: Each worker node has a Docker container running, which hosts the configured pods.
Pods: A pod is a collection of one or more containers that operate logically as a unit on a node.

⌚ Describe namespaces. Why is using a single default namespace problematic?

With the use of namespaces, you may divide your cluster into virtual clusters so that you can organise your applications logically while keeping them totally independent from the other groups (so you can for example create an app with the same name in two different namespaces).

Over time, it becomes challenging to keep track of all the apps you administer in your cluster when utilising the default namespace alone. With namespaces, grouping apps into sensible groupings is made simpler. For example, a namespace for all monitoring applications, a namespace for all security applications, etc.

Additionally, namespaces can be helpful for managing Blue/Green environments, where each namespace can have a unique version of an app and share resources with other namespaces (namespaces like logging, monitoring, etc.).
One cluster, several teams is another application for namespaces. It's possible for teams using the same cluster to step on each other's toes. Because there can only be one app with the same name in Kubernetes, if they wind up generating an app with the same name, it signifies one team has replaced the app from the other team (in the same namespace).

⌚ What happens when a master fails? What happens when a worker fails?

Kubernetes is meant to be resilient to any master or worker node failure. The cluster's nodes will continue to function if a master fails, but until the master is back up, no modifications can be made, such as adding or removing service members. The master no longer receives signals from a failing worker. The node will be labelled as NotReady if the master does not receive status updates from the worker. The master moves all running pods from a node that is not ready for five minutes to other nodes that are.

⌚ Describe a DaemonSet?

When you need to execute one or more pods on all (or a portion of) the nodes in a cluster, Kubernetes uses daemon sets. A DaemonSet is typically used for monitoring and logging on the hosts. To push health or log data to a centralised system or database, for instance, a node needs a service (daemon).
As the name implies, daemon sets can be used to execute daemons (and other tools) that must run on all cluster nodes. For example, Quobyte, Glusterd, Ceph, and other cluster storage daemons; Fluentd or Logstash; or Monitoring Daemons (e.g. Prometheus Node Exporter, collectd, New Relic agent, etc.)

⌚ What do you mean by Operators

"Operators are K8s software extensions that manage applications and their constituent parts by using special resources. The control loop is a key notion that operators adhere to."

⌚ Why are Operators necessary?

In contrast to managing stateless applications, where achieving the desired status and upgrades are both handled in the same way for each replica, managing applications in Kubernetes is more complex. Due to the stateful nature of stateful applications, upgrading each replica may require distinct handling because each copy may be in a different status. As a result, managing stateful applications frequently requires a human operator. This is what Kubernetes Operator is designed to help with.

Additionally, this will assist in automating a routine procedure across several Kubernetes clusters.

⌚ Describe GKE?

GKE, short for Google Kubernetes Engine, is a tool for managing and coordinating Docker container systems. The container cluster can also be orchestrated with the aid of Google Public Cloud.

⌚ How can Kubernetes be run locally?

The Minikube utility can be used to install Kubernetes locally. On the PC, a single-node bunch is running. As a result, it provides the ideal method for users who are still learning Kubernetes.

⌚ What is load balancing in Kubernetes?

One of the most popular and accepted methods of exposing the services is load balancing. In K8s, load balancing can be done in one of two ways:
Internal load balancer: This kind of balancer distributes the required incoming load to the appropriate pods while automatically balancing loads.
The traffic from external loads is routed to backend pods via an external load balancer.

⌚ What do Kubernetes labels mean?

Labels are a group of keys that each hold a few values. Replication controllers, pods, and related services are all linked to the key values. Typically, labels are added to an item at the time of creation. Users have the ability to alter them in real time.

How can API security be provided on Kubernetes?

On Kubernetes, there are several techniques to ensure API security.

use the appropriate authentication mechanism with API server authentication mode = Node.
Make kubeless such that authorization-mode=Webhook secures the API.
Make certain that the Kube-dashboard adheres to a tight RBAC (Role-Based Access Control) policy.

Consider a multinational corporation that is working to find a solution to the monolithic code base problem and has a highly dispersed system. How do you believe the business can resolve its issue?

They can change their monolithic code base to a microservice design to address the issue, and each and every microservice can then be viewed as a container. Therefore, Kubernetes may be used to deploy and manage all of these containers.

Let's say a business decides to implement new technologies in order to distribute its tasks more efficiently. How can the business efficiently distribute its resources?

Kubernetes is the only tool that can solve this issue. Kubernetes ensures that resources are used just as needed by that specific application and that they are efficiently optimised. Therefore, the organisation may efficiently distribute its resources by using the best container orchestration technology.

Let's say a business wishes to update its deployment strategies and develop a platform that is considerably more flexible and responsive. What are the best ways this business can accomplish this to please their clients?

The organisation requires a platform that is scalable and responsive in order to swiftly transmit data to the client website and provide millions of customers with the digital experience they would expect. To accomplish this, the business should switch from its proprietary data centres, if any, to a cloud environment like AWS. Additionally, they should put the microservice design into practise so that they can start utilising Docker containers. They can begin using Kubernetes, the greatest orchestration platform available, once the basic framework is ready. This would allow the teams to build applications independently and deliver them extremely quickly.

What distinguishes a replication controller from a replica set?

The functions of a replication controller and a replica set are very similar. They both make sure that a certain number of pod replicas are active at all times. Selectors are used to replicate pods, which is where the difference lies. Replication controllers employ Equity-Based selectors, whereas replication sets use Set-Based selectors.

Equity-Based Selectors: This kind of selector enables label key and value filtering. So, to put it simply, the equity-based selector will only search for pods that contain exactly the same word as the label. Using this selection, you can only check for pods whose label app is identical to nginx, for instance, if your label key is app=nginx.

Selector-Based Selectors: These selectors enable the filtering of keys based on a set of values. Thus, the selector-based selector will search for pods whose label has been referenced in the collection, to put it another way.
Consider this: Your label key may read "app in" (nginx, NPS, Apache). The selection will treat your app as a true result if it is equivalent to nginx, NPS, or Apache in this case.

What is resource monitoring for containers?

For users, it is crucial to comprehend how well an application performs and how resources are utilised at each abstraction layer. Kubernetes has accounted for this by factoring in the management of the cluster by creating abstraction at various levels, including container, pod, service, and entire cluster. Now that each level can be watched, container resource monitoring is being done.