Autok8s is assignment rquire to be done with in 1 week

Tasks requirements:

• Tasks requirements:
  • Automate the deployment of a local kubernetes cluster
  • Automate the deployment of a testnet cosmos rpc node
  • Automate the deployment of grafana and prometheus to view the resources used such as cpu / memory / disk space etc..
  • Estimated time: 3-4hours.
  • Duration: 1week.

Solution plan

• Use kubelet=1.28.2-1.1 - kubeadm=1.28.2-1.1 - kubectl=1.28.2-1.1 and Ubuntu22.04
• Install promethues
• k8s-cosmos-gaia | I build a gaia image.

Cluster Description

• This version deploys a single-node cluster with kubeadmin. If needed, it can be modified to a high-availability cluster with k8s source.

Role	Quantity	Description
Deployment Node	1	Runs ansible commands, reused as the first master node
etcd Nodes	1	Note: reused as master nodes
Master Nodes	1	A cluster with 1 master nodes;4c8g;
Node Nodes	2	Nodes running application workloads; machine configuration can be upgraded/increased as needed; 8c16g for cosmos

How to use it

• Prepare your running env. Now we setup with 1 master node and 2 worker nodes.
• Set up passwordless SSH login

$ sudo ssh-copy-id $IP

• Git clone this repo on master node. It also used as ansible runner.

$ git clone https://github.com/terrydevops/autok8s.git

• Install Ansible

$ sudo apt update

$ sudo apt-get install ansible

$ cd ./autok8s

• Change the IP list in inventories/group_vars/all.yml

• run ansible-playbook sudo ansible-playbook playbooks/setup_kubernetes.yml

Cosmos is read once the ansible playbook done

Grafana and prometheus

For gaia-0 cpu/memory

Grafana dashboard http://54.90.137.176:30300/dashboards username:admin password:admin

Todo list

• Run ansible with runner
• Timezone Sync (chrony)
• Test cases
• Support HA with master node
• Support HA with etcd
• Support different network: calico, cilium, flannel, kube-ovn, kube-router
• Upgrade/remove support
• Build k8s from source code
• Full env for cosmos peer to peer
• Other requirements(kubernetes-sigs/metrics-server load-balancer docker runtime )