Consider organization

[simenandre]

As we evolve our projects, we should carefully consider how we'd like to architecturally extend the project.

Summary

• Core API is deployed to App Engine (project: ayr-core-service)
• Core Backoffice is deployed to App Engine (project: ayr-core-service)
• pgadmin is deployed to Compute Engine (a default VM, with Docker Container, on a separate project)
• Workspace Agent is deployed to Cloud Run (uses Pub/Sub and Cloud Scheduler, separate project)
• Onboarding is deployed to App Engine (on a separate project)
• User API is deployed to App Engine (on a separate project)
• ALSO Pricelist Sync is deployed to Cloud Run (uses Pub/Sub and Cloud Scheduler, on a separate project)
• EPP Outline server is deployed to Compute Engine (on a separate project)

Things to consider

• What is a good organization for our kind of organization? We should read what Google thinks. We should also think about the maximum projects we can have on each billing account.
• Logging agents currently needs to be deployed to each project.

[simenandre]

👆 Current setup

@bjerkio/thebranches has access to this file here:
https://drive.google.com/file/d/1KSaDZuKfCen0gAH4elIBYou9a8bW2Qzu/view?usp=sharing

[simenandre]

A year has passed since this was last updated, and the more time that passe,s the more I feel we should refactor things.

We have grown to use a Kubernetes cluster now as well.

[simenandre]

• Core API should be move to Kubernetes
• Core Backoffice should be move to Kubernetes
• pgadmin should be killed
• Workspace Agent is being modernized
• Onboarding should be moved to Kubernetes (since it's a Next.js, actually using the backend)
• User API should be killed
• ALSO Pricelist Sync should be killed
• EPP Outline server should be killed (use a fixed internet with normal VPN?)

[simenandre]

Just killed Kubernetes temporarily. Seems dumb to spend money on it when we're not really using it.

[simenandre]

I'm closing this as done and successful now :)