I’d like to host a C++ project on GitLab. It needs to compile on Windows (x86_64), Mac(x86_64 and arm64) and Linux (x86_64). I have docker containers that already properly host the development environments for Windows and Linux. These should be able to run on an EC2. We already have cloud hosted Mac Minis to handle the mac side of things but it’ll need to be outside of AWS.
In a perfect world, what I’m looking for is the ability to spawn cloud runner instances to handle the compiles each time the pipeline is run. Once the pipeline is complete, the instances should go away.
I read the docs on AWS EC2 Autoscale and that seems like what I want (for Windows/Linux) except I saw no obvious way for it to differentiate the different platforms and container images. It seems like it can only spawn the same image anytime there’s a job.
Has anyone had success in setting up this type of complexity?