Kubernetes

Before we deep dive into Kubernetes we need to understand what is a Container (with Docker) , why we use it.

What is Docker Container

Containers are a type of software that can virtually package and isolate applications for deployment. It shared the underlying OS and resources rather than running it’s own

• Logical isolation of underlying OS Resources like CPU , RAM ,Network

• Provides Process Isolation , Network Interface , IPC , Filesystem Mounts (Data volumets) and Isolation on Kernel

• We should be aware that , Container
  • Provides service isolation between containers (inherently processes and software packages)
  • Provides consistent Runtime with Low Overhead
  • Container have limited resource
• Containers are Running instance of an Image
  • An image is a lightweight, standalone, executable package of software that includes everything needed to run an application: code, runtime, system tools, system libraries and settings.
• Read more on Containers on
  • Google Article
  • TechTarget Blogs

Some Mictoservice Architecture Patterns

It’s essential to understand below patterns as K8s has been built to natively support it.

• Scalability , Scaling of Process (Horizontal Scalling) as needed
• Availability , Automatically recover from faliure of service/process
• Service Discovery , How a service discover another service’s instance which is healthy
• Reliability , Revover from service failures and is enabled by Health Check , Self Healing and most importantly Automatic Rollouts
• Secret and configuration management , Drive all the changes and properties as Cnfiguration code
• Observability , some aspects are supported nativily namely Health Check at least in a Primitive way
• And many more…

Application Architecture on Kubernetes

How Application Architecture is comparable to VM as well as Netflix based Microservice

“Processes on VMs” vs “Microservice on K8s”

“Microservice on PCF with Netflix OSS” vs “Microservice on K8s”

• K8s provides Platform resources (Container,Service,Configuration,Volume Mounts,..) as a Service and it can be enabled by calling API and defining Object Spec
• Benefits of using K8s
  • Workload Scalability by providing easy Horizontal infrastructure scaling , Auto-scaling
  • Decoupled microservice architectures by providing Namespace , Service and Pod/Container isolation
  • Abstraction of Infrastructure by allowing custom Ingress Controller and Load Balancer
  • Self Healing Systems
  • And more… DNS management, resource monitoring, Secrets Management , logging and storage orchestration

How to provide Object Spec

• Object Specs are described via a yaml file
• Each Specs should have apiVersion , type and name along with the desired state Specification
• Kubernetes takes care of making necessary changes to achieve that desired state
• An yaml file can be run to create the K8s object via kubectl apply -f file_name.yaml
• A sample yaml definition file looks as below

apiVersion: v1 #K8s API version. (mandatory)
kind: Pod #K8s object type. (mandatory)
metadata:
  name: example-app #Name of the Pod.(mandatory) Each resource's unique identifier
spec: #Detailed specification for Resource kind (mandatory) 

Namespance

Before we start using K8s we need to have our own Namespace on the Cluster

• Namespaces are a way to divide cluster resources between multiple users/teams
• Each Namespance have allocaed quota on Memory and CPU. Also have certain permission on what K8s Resorces it can create/use
• Resources in namespace can be found by kubectl api-resources --namespaced=true