jinwoo1225 / cloud_kubernetes

Kubernetes, Node Exporter, Grafana, Prometheus, EFK stack, helm chart skill stack practice with comento

cloud_kubernetes(Comento Bootcamp 2021.09~10)

How to ?

# copy it to your repository
git clone https://github.com/jinwoo1225/cloud_kubernetes.git

week 1

• install kubernetes on ubuntu 20.04 LTS (AMD64, ARM64)

  # amd64 arch machine
  bash ./install_kubernetes_ubuntu_amd64.sh
  
  # arm64 arch machine
  bash ./install_kubernetes_ubuntu_arm64.sh

  bcz im using m1 mac 🤣

• creating kubernetes cluster

  ## create cluster with your custom cidr. 
  ## https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/create-cluster-kubeadm/#initializing-your-control-plane-node
  sudo kubeadm init --pod-network-cidr yo.ur.ci.dr/idk
  
  ## copy config to your storage
  mkdir -p $HOME/.kube
  sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
  sudo chown $(id -u):$(id -g) $HOME/.kube/config

• deploy calico CNI

  ## download calico deployment file
  curl https://docs.projectcalico.org/manifests/calico.yaml -O
  ## apply it to your cluster
  kubectl apply -f calico.yaml

• deploy grafana

  if you are deploying it to Master Node, you should untaint it

  kubectl taint nodes --all node-role.kubernetes.io/master-

• (advanced) deploy grafana with service, persistent volume(PV), persistent volume claim(PVC)

  check : grafana.yaml

• (advanced) access grafana dashboard

  curl localhost:3000

week 2

• create simple nodejs application

  check : simple-webapp

• build application as container (arch : linux/amd64, linux/arch64, linux/arm/v7) and push to hub.docker.com

  to build multi platform container check : buildx

  Why I'm using it ? check : meetup.toast

  # this command will build simple webapp to support amd64, arm64, armv7 and push to hub.docker.com
  sudo docker buildx build --platform=linux/amd64,linux/amd64,linux/arm/v7 -t jinwoo17962/simple-webapp ./simple_webapp --push

• deploy application to kubernetes cluster

  check : nodejs.yaml

• (Advanced)

  • create simple socket transaction program for cluster

    • check : python-socket-client

      it will create replica of 4 and send signal to server

    • check : python-socket-server

      it will create thread to each client, and send success to each valid request

  • build it, push it, deploy it, test it

    sudo docker buildx build --platform=linux/amd64,linux/amd64,linux/arm/v7 -t jinwoo17962/python-socket-client ./python_socket_client --push
    
    sudo docker buildx build --platform=linux/amd64,linux/amd64,linux/arm/v7 -t jinwoo17962/python-socket-server ./python_socket_server --push
    
    kubectl apply -f socket_client.yaml
    
    kubectl apply -f socket_server.yaml
    
    # logs with -f option will stream the log
    kubectl logs deployment/socket-server -f

week3

Install Prometheus, Node Exporter, Grafana in my cluster. (namespace : monitoring-prom)

Create a namespace

create namespace object to simplify maintaining a applications.

Deploy Node Exporter

deploy as daemonset to ensure every nodes have a single pod instance.

set hostNetwork, hostIPC,hostPID, hostPath(as readonly) to monitor node's metric.

Deploy Prometheus

• Persistent Volume and Persistent Volume Claim to store logs

• ConfigMap Use Service Discovery to find every node that contains node-exporter

• ClusterRole, ClusterRoleBinding, ServiceAccount To use Service discovery (ClusterRole, ClusterRoleBinding)

• Deployment, Service https://prometheus.io/docs/prometheus/latest/getting_started/ no-lockfile option will help pods to rolling update if not setted, two prometheus pod will get one storage and get deadlocked at that moment.

  args:
    - "--config.file=/etc/prometheus/prometheus-settings.yaml"
    - "--storage.tsdb.path=/prometheus"
    - "--storage.tsdb.retention.time=6h"
    - "--storage.tsdb.no-lockfile"

Deploy Grafana

Using week1's grafana.yaml to install :)

to add database you will have to use URL {prometheus-service name}:{portname} to your Data Sources.

example : http://prometheus-svc:9090

Deploy EFK

Deploy Namespace

• namespace : logging-elas

Deploy Fluentd

• ServiceAccount, ClusterRole, ClusterRoleBinding, Daemonset

Deploy ElasticSearch

• Namespace
• Service
• Persistent Volume, Persistent Volume Claim You should not use Persistent Volume in production. you should checkout dynamic provisioing
• Stateful Set

Deploy Kibana

• Service
• Deployment should contain enviroment variable called ELASTICSEARCH_URL to be serviceURL ex) elasticsearch-svc:9200

week4

Install helm chart to your System

bash ./week4/install_helm_amd64.sh

Metric Monitor(Node-Exporter, Prometheus, Grafana)

create chart template

helm create metric-monitor

will create template and change the value in values.yaml

create values for applications : example

helm install metric-monitor . -n metric --create-namespace

• -n metric : use metric namespace
• --create-namespace : create namespace if required namespace is not present.

Log Monitor(ElasticSearch, Flunetd, Kibana)

helm create log-monitor

helm install log-monitor . -n logging --create-namespace