Microservices and Container Orchestration

Microservices Architecture

• Microservices architecture is often built using RESTful web services.
• Components are independently scalable, meaning you can adjust one part of the system without affecting others.
• This leads to a smaller code surface area.
• Reduces dependency between application modules.

Deploying an Application with a Container

• Container solutions are consistent, lightweight, and portable.
• They can run anything, anywhere.

Kubernetes and Docker

• Docker stores images.
• Kubernetes Engine contains components like Network, Etcd, Kubelet, and Master.
• Multiple nodes work together in one Docker.

Official Definition of Kubernetes (k8s)

• Kubernetes stems from the Greek word "Helmsman," referring to a ship's captain.
• It's an open-source container orchestration framework, developed by Google.
• Kubernetes helps in managing containerized applications across various deployment environments (physical, cloud, virtual).

Need for Container Orchestration Tools

• Increased container usage created a demand for a proper way to manage them.
• Features offered by these tools include high availability (no downtime), scalability, and disaster recovery (backup/restore). Master nodes control slave/worker nodes.

K8s as the Operating System of the Cloud

• Kubernetes (k8s) is the de-facto platform for managing and deploying cloud-native applications.
• It abstracts cloud resources and schedules application microservices.
• It standardizes disparate environments (private vs. public clouds).
• It represents a critical step toward a hybrid cloud solution.

Quick Analogy

• Think of sending goods via a courier service: packaging and labeling the goods.
• Kubernetes packages an application into a container and provides the necessary configuration.

Kubernetes Architecture Diagram

• Master nodes manage k8s processes.
• The API server is the entry point to Kubernetes.
• The controller manager tracks cluster activity.
• The scheduler ensures Pod placement.
• Pods (units) run on worker nodes.
• ETCD is the underlying storage layer. Workers are physical or virtual machines. Kubelet is the primary node agent. Applications run on worker nodes.

Master Node vs Worker Node

• Master nodes handle the primary processes, critical for Kubernetes function.
• Worker Nodes have a large workload, and significant resources, to run the application.

Packaging Apps for Kubernetes

• Applications need to be packaged as containers, wrapped in Pods, and deployed using declarative manifest files.

Declarative Model and Desired State

• Kubernetes relies on the declarative model and desired state concept.
• Applications define their desired state in manifest files, which are sent to the API server.
• The Kubernetes cluster implements the specified state, ensuring consistency.
• A controller assures the actual state matches the desired state.

Getting Kubernetes

• Three common ways to access Kubernetes include: Playground, Hosted Kubernetes, and DIY install.
• Several Cloud providers offer Hosted Kubernetes solutions
• AWS EKS
• Azure AKS
• Linode LKE
• DigitalOcean DOKS
• Google GKE

kubectl

• kubectl is Kubernetes' command-line tool.
• It translates user commands into HTTP REST requests using JSON. Configuration files define clusters and give authorized users access
• A configuration file (kubeconfig) specifies the cluster and API server.

Sample kubeconfig File

• Examples of a kubeconfig file show various sections, including clusters, users, and contexts.

Pod

• The smallest unit in Kubernetes.
• It abstracts a container.
• Allows a single or multiple containers per Pod.
• Each pod has its unique IP address. IP addresses are re-assigned on Pod recreation.

Pods and Shared Networking

• Pods share networking resources.
• Multi-container Pods share the same IP, ports, and routing table.
• Single-container Pods have full access to their own resource block

Pod Network

• Pods can directly communicate with one another in the flat network, eliminating complex port mappings.

Pod Immutability

• Pods are immutable.
• Modifications are not conducted on existing Pods. When updates are needed, existing Pods are replaced, not modified.
• Failures trigger a replacement with identical new Pods.

Pod Manifest Files

• YAML files define Pods. The YAML file defines characteristics and properties.
• kind specifies the object type (e.g., Pod).
• apiVersion outlines the schema and API version used

kind

• Defines the type of object being described. Pods, services, deployments, persistent volumes, storage classes, namespaces, secrets, and configmaps are examples of commonly defined kinds of objects.

apiVersion

• Format is /. StorageClass example is storage.k8s.io/v1.
• Pods are part of the core API group.
• Documentation for Kubernetes API is available online.

metadata

• Attributes like names, labels, annotations, and namespaces are in this section.
• Names help identify objects within the cluster. Labels let you categorize objects. Annotations provide a space where you can attach extra data for specific purposes. Namespaces partition the cluster for logical groupings.

spec

• This section describes the structure of resources. Contains data about containers running inside the Pod.

kubectl commands - Pods

• kubectl apply deploys Pods.
• kubectl get pods lists running Pods.
• kubectl describe pod <pod-name> displays detailed Pod information.
• kubectl logs <pod-name> prints logs from the specified Pod.
• kubectl exec <pod-name> gives you (command line) access to the container within the specified Pod

This quiz explores key concepts in microservices architecture and container orchestration, including the use of Docker and Kubernetes. Understand how independent scalability and reduced dependencies benefit application design and deployment. Test your knowledge on the fundamentals that support modern software development.