Google Kubernetes Engine Demo

Example of github-action that automatically deploy flask server to Google Kubernetes Engine

Demo Video

I made a video for this demo using chinese.

Main Purpose

• Simple flask server with simple test just for demo
• Github action to automatically run unit test on each push and pull request
• Github action to automatically deploy to google cloud container registry after each tag push
• Github action to automatically deploy to google kubernetes engine after each tag push

Directory Structure

• Program
  • main.py: route
  • core.py: core logic
  • test.py: test for core logic
  • requirements.txt: dependencies
• Environmet Variable
  • .dockerenv: local docker
  • .envrc: local shell
  • containers.env of .gcloud_deployment.yaml: gke
• Settings of ignore files
  • .gitignore: ignore files of git
  • .dockerignore: ignore files of docker
• Others
  • Dockerfile: script of docker
  • gcloud_deployment.yaml: deployment and service of gke
  • .github/workflows/workflow.yaml: script of github action
  • Makefile: some handful script, readme.md

Google Cloud Settings

1. For test just create gke cluster with only one node(g1-small) at us-central1-a
2. Add a service account with these IAM
   • Kubernetes Engine Developer
   • Service Account user
   • Cloud Storage Admin
3. Generate key from this service account and
   • Run cat service_account_key.json | base64
4. Set google cloud secret to github
   • secrets.GCP_PROJECT_ID: Your google cloud project id
   • secrets.GCP_SA_KEY: The base64 code generated from service account key from previous step

• Step 2, 3, 4: To grant gke deployment privilege to github action.

Github Action Settings

See .github/workflows/workflow.yaml

• jobs.test runs unit test on every push and pull request (including tag push)
• jobs.deploy runs deploy to container registry and runs deploy to kubernetes engine on every tag push

Development

• Add route at main.py, Add core logic at core.py, Add unit test for core at test.py
• Run make test for unit test
• Run python app.py to start local test server
• Run curl "http://127.0.0.1:5000/entry?str1=hello&str2=world" to test server on local
• Git add, commit, push to trigger CI to run unit-test
• Git tag v0.1, push tag to trigger CI to run unit-test and deploy to container registry and gke
• Run curl "https://PRODUCTION_URL/entry?str1=hello&str2=world" to test server on production