In Microservice Architecture, we have many small services as repositories and the released time is very short. It's very high-load if we do all thing manually.
The jobs including:
- autotag a repo when there are new code in the master branch
- docker auto push image with new version
- deploy the service in ec2 using docker
- add .gitlab-ci.yaml in a service's repo
- add code
stages:
- tag
- build
- deploy
include:
- remote: https://github.com/nvhoc/gitlab-jobs/releases/download/v1.0.0/default.gitlab-ci.tag.yml
- remote: https://github.com/nvhoc/gitlab-jobs/releases/download/v1.0.0/default.gitlab-ci.docker.yml
- remote: https://github.com/nvhoc/gitlab-jobs/releases/download/v1.0.0/default.gitlab-ci.ec2-deploy.yml
- define CI&CD variables:
3.1. to include .gitlab-ci.tag.yml:
- generate a pair of key ssh by ssh-keygen => private_key && public_key
- set SSH_PRIVATE_KEY=${private_key} (text)
- add public key in "Deploy Keys" at "CI&CD settings"
- set GITHUB_EMAIL=${your bot email}
- set GITHUB_USER=${your username}
3.2. to include .gitlab-ci.docker.yml:
- set DOCKER_USER_NAME=${your bot username in your docker registry}
- set DOCKER_PASSWORD=${your bot password in your docker registry}
3.3. to include .gitlab-ci.ec2-deploy.yml:
- note: the file support for simple deployment with docker-compose. I will add new file for k8s later.
- set EC2_PEM_KEY=${content of pem file}
- set EC2_USER=${ec2 username}
- set EC2_HOST=${ec2 public address}
- set EC2_DEPLOY_PATH=${ec2 path store docker-compose deployment} find a example at ./example/docker-compose
3.4. to include .gitlab-ci.yarn-build.yml:
- set BUILD_PATH=${path of folder from building process}
3.5. to include .gitlab-ci.s3-sync.yml:
- set BUILD_PATH=${path of folder that you want to sync to s3}
- set S3_DEPLOY_BUCKET=${s3 bucket}
- set AWS_ACCESS_KEY_ID=${aws access key}
- set AWS_SECRET_ACCESS_KEY=${aws secret key}