github action notes

ci / integration :

build

test 

merge

cd / delivery :

auto release to repository

cd / deployment :

auto deploy to production

=====

ci = pass test

cd = deliver code to production

======

## gitactions

1 create workflow = superlinter.yml

// linter = run check codes that meet specific criteria or standard

=======

## workflow

terdiri dari 5 :

1 event

2 jobs 

3 runner 

4 steps

5 actions

1 event = trigger for workflow 

- example: when someone push new code 

// contoh event di yaml file :

on: push 

^ bakal jalanin isi dari jobs 

2 jobs = jobs todo 

jobs:

super-lint:

    name: Lint code base

    runs-on : ubuntu-latest         // runner => container 

    steps:

- name: checkout code

    uses: actions/checkout@v2        // check our code

- name: Run Super-Linter

  uses: github/super-linter@v3      // run linter 

  env:

    DEFAULT_BRANCH: main

3 runner = container yg buat jalanin obs

runs-on : ubuntu-latest 

## by default github runs our code in github environtment 

## ada ubuntulinux / windows / mac 

4 steps = declare berapa jumlah step yg akan dijalankan 

stepnya jalanin linter**

=========

superlinter support multiple linter ( understand multiple language )

===========

awal bikin workflow penamaan penting

mygitactions/.github/workflows/superlinter.yml     // in main 

## trs di commit ke main branch

=========

## msk ke <> code 

## check status icon 

check pass = turn green 

check fail = turn red

atau ke actions tab 

=============

CD

klo udah ci pipeline selesai dan image dipush ke registry saatnya bikin cd.

create new repo.

isinya config repo :

- kubernetes manifest / docker compose

push ke test/QA env --> dimonitor pake prometheus / grafana 

prometheus = script metric from app

grafana = qa bs visualize change, check latency, check kpi 

biasanya ada 3 cd stage:

test/qa

staging  : pretend as production. testing live stage

production : customer access 

==============

push model:

registry -> dipush ke 3 stage ( kubectl apply ) -> dimonitor pake prometheus / grafana

pull model:

argoCD diinstall di tiap kubernetes cluster 

argoCD = cek state repository, cek state masing2 cluster ( di pull ) 

klo di cluster ada perubahan lgsng ngecek dan ngasih tau klo out of sync

bs otomatis di sync sama si agro atau manually triggered

otomatis sync test/qa env dengan staging env via argo CD

fungsi argoCD : sync dan ngecek state

============

// argo rollout canary deployment

user --- loadbalancer --- app ( kubernetes pod ).

90% traffic

user --- loadbalancer --- app v1

|

|        10% traffic

--------- app v2

90%

10%

80%

20%

**mindahin traffic ke versi baru sampe smuanya pindah ke versi release yg baru

podnya** jg dari aplikasi yg lama dipindah ke app yg baru 

==============