K8s, Helm, Rancher

This is a documentation of my path in exploring, experimenting and learning Kubernetes (K8s), Helm, and Rancher.

Introduction

• follow https://kubernetes.io/docs/tutorials/
• ways of running k8s: minikube, kind, kubeadm
• start with minikube for local testing
• kubectl to communicate with k8s cluster
• deploy fullstack app to local k8s cluster

notes k8s:

• deployment: manages an instance of an application, ensures desired state inside the cluster
• service: exposes deployment to external network
  • allows for pods to die and be recreated without affecting access to the application because it routes traffic and abstracts away the pods
• pod: group of one or more application containers, it includes shared storage, IP address and information about how to run.
  • Containers should only be scheduled together in a single Pod if they are tightly coupled and need to share resources such as disk.
• configMap: used to store non-confidential data in key-value pairs.
  • as volumes updates on pull
  • as env vars only updates on restart of pod

notes minikube:

minikube needs a tunnel to expose services with type LoadBalancer

Testing Service Communication

The Goal is to Measure and compare service communication in Kubernetes using two approaches:

• Services in separate Pods communicating over the network via a ClusterIP Service.
• Service containers in one Pod communicating via the filesystem.

Visualization: Figma

To Measure this the data for benchmarking will be doubled for every iteration.

First we record the time it takes until the communication has succeeded.

Later maybe add some more metrics:

• latency
• throughput
• cpu
• memory
• i/o
• bytes transferred

Also interesting would be to look into what happens during failures, scaling and rollouts.

##questions

What data will I use?

• randomly generated data
• format:
• size steps: 64 B, 1 KiB, 2 KiB, 4 KiB, 8 KiB, ... … 512 KiB, 1 MiB, 2 MiB, 4 MiB, 8 MiB, 16 MiB

When does the time measurement start?

Validity of data:

• timing on different nodes might drift

How will it work: Network:

• record t0
• send data via http post + t0
• recieve response
• record t1
• calculate t1 - t0

Files:

Common issues: image not found: eval $(minikube docker-env)