Updating thousands of images at scale

General Discussions General
1

Hi,

So we are deploying our client’s application using images. Basically our use case

  1. There’s “platform-image” docker image which we create with the platform libraries/framework
  2. And create “user-image” docker image which uses “platform-image” as a base and build on top of it the “user-image”.

So when we release platform-image we’d like to update all user images whichcan are thousands of user images.

The ideas we have so far:

  • Use a sidecar container in Kubernetes. That means that we require 2 images and make a hard dependency on k8s (e.g cannot run locally for debugging without k8s)
  • Have the platform libraries being installed upon a start-up of the image.

Do you have any other suggestions/ideas?

2

I guess no suggestions or is the question is a bit unclear?