Overview

App

Auth, base setup and endpoints

• Auth0
• /health
• /metrics
• Dockerfile

CI

Build and test pipelines.

• cloudbuild.yaml
• vob.yaml

Resources

Provisioned services

• DB, PubSub via tf-wonka
• secrets via Secret Manager

Deployment

Runtime configuration & workloads.

• helm-chart-values
• secrets / env vars
• file mounts / volumes

CD

Continuous delivery and sync

• ArgoCD
• pipeline deploy

1. App

   If you start thinking about the deployment already during the development, it can simplify the process a lot.

   We have some app requirements that are good to check before starting the deployment process. Satisfying them will help the process to be nice and smooth.

   You can also use the Dockerfile templates instead of writing it from scratch.

2. CI

   The next step is to package the app into the image. Currenlty, we support only Cloudbuild to do that.

   We encorage you to use the templates and not write cloudbuilds from scratch every time. There’s no preference between vob and buildx — use whatever you like more.

   You can configure cloudbuilds for different events:

   • push — push to a branch
   • pr — pr is created
   • master — merge to master
   • stable — merge to stable
   • tag — tag is created
   • nightly — regularly runs at night

   Once you have a file(s), write a message to the #platform-infrastructure Slack channel with the name of the repos and triggers you want to add. Also, add robot-infra GitHub user as an admin to your repo.

   If you don’t want to wait for us to create the cloudbuild for you, you can do it yourself in the tf-wonka repo. Here’s the example

3. Resources

   If your app is using any resources, like a database or a pub/sub — now is the time to create it. You need to go to the tf-wonka repo and define them. Once you’de assign infra team to be a reviewer of your PR.

   If you don’t will comfortable doing terraform — drop by the infra room, we’ll help you out to feel better with it.

4. Deployment

5. CD