Cache for runner worker with spot instances

[quezadasubiabre]

I need some help. I hope someone can help me.

I am implementing the following scenario GitLab CI docker-machine runner - one runner agent. I would like to have in the runner worker a cached dokcer image, but as I am using spot instances, every time a new instances start the image is not there and is pulled from internernet (ECR in my case). This image is a little bit heavy and it takes 6 minutes to dowload the image.

I was trying to create an Auto Scaling group for the worker runner. So in that way I could use a warm up pools and have an AMI with the base docker image already downloaded, but he configuration accepted for the worker runners is very simple

What do you recommend??

Thanks in advance