Kubernetes (k8s) is a container orchestration tool introduced by Google in 2014, designed for deploying and scaling applications, especially microservices. It can run multiple container runtimes and offers various alternatives for learning and production, such as Minikube, K3s, and Microk8s, each tailored for different use cases and resource requirements. Pods, the smallest deployable units in Kubernetes, can be created through deployments and can leverage features such as replication and management through a control plane.

1. Kubernetes

2. Kubernetes (k8s)
• Tool to deploy and run Applications in containers
• Google in 2014
• Container Orchestration and scaling
• Best Suited for Microservices and Distributed Systems
• Why not Docker?
• Runs largely on a single server – Single Point of Failure
• Docker Cluster and Swarm never became popular
• Support for other container Runtimes – Kubernetes can run multiple
container runtimes

3. Kubernetes (k8s)

4. Kubernetes (k8s)

5. Kubernetes Getting Started
• Play-With-K8s
• Miniqube – Learn on own system
• No Separate Master/Worker Nodes
• Single System is both Master and Node
• KubeAdm
• Reqires multiple VMs

6. Alternatives to MiniQube
• Minikube, while excellent for simple, local Kubernetes environments,
is NOT recommended in a production environment.
• Scalability Challenges – best suited for learning the basics of
Kubernetes

7. Alternatives to MiniQube
• KIND - Kubernetes in Docker
• Fast and lightweight container creation
• Declarative approach to cluster Configuration similar to Terraform or other
IAC
• Multiple Kubernetes Version support
• Again best suited for learning and development support
• Harder to setup over MiniQube

8. Alternatives to MiniQube
• K3S

9. Alternatives to MiniQube
• K3S
• Both Single and Multiple Node Configuration
• Minimal Memory footprint
• Inbuilt Security Support – TLS, RBAC, Reduced Surface
• Containerd (Similar to Docker)
• Best Suited for Edge and IOT

10. Alternatives to MiniQube
• Microk8S
• Single node like MiniQube
• Less Resources
• Multiple Plugins
• More Secure
• Closer to Production – so easier to leverage going from Development to
Production
• Steep Learning Curve
• Better Integrated with CI/CD tools
• Private Cloud and OnPrem Support

11. Alternatives to MiniQube
• Rancher
• Production Ready Capabilities – Security, Multi-Node, Scaling
• Multi-Cluster Management
• Resource Heavy
• Steep Learning Curve
• Hybrid and Multi-Cloud Support

12. PODS
• A group of 1 or more container
• Smallest deployable - Similar to what we have seen with docker-compose
• Created via Deployments
• StateFulSet resource for managing state
• Can be horizontally scaled – Called Replication in Kubernetes world.
• Name of a pod is a valid DNS subdomain value – This allows them to be
accessed via name (Like docker compose does for services)
• PodTemplates are like Docker Images and are specifications for creating
Pods

13. PODS
• Static PODS are managed by the kubelet Daemon or specific node

14. Kubernetes
• WSL2
• Docker Desktop
• Ubuntu 18 from App Store
• Terminal

15. Kubernetes
• Kubectl
• Minikube
• K3s
• Microk8s
• Kind
• Rancher
• Docker Desktop

16. Kubernetes
• Container Orchestration
• Automating SW Development
• Scaling
• Management
• Cloud Native Computing Foundation

17. Kubernetes
• Primary/replica architecture
• Control Pane
• ETCD = Key value data Store - Configuration data of the cluster and
distributed
• Consistency over availability
• API Server
• Scheduler