Multi instance deployment, with custom plugins and staging. Devops


I am a Java software engineer, I have never used node. I want to try this system for a large company.

I am looking to deploy reaction on AWS EB multi instance with a docker image, and external database.
My deployment pipeline works by pulling a git repo, building/packaging it and then uses the AWS CLI script to send the package to EB to deploy.

So i have created a new git repo and then on my desktop i cloned that and ran ‘reaction init’ inside it. But there is nothing to push because gitignore ignores is set to ignore everything. So my deployment pipeline has nothing to pull.

What is the solution to make a normal devops scenario with reaction commerce?

I need to add a lot of custom plugins, so i don’t just want to pull the reaction commerce repo to my servers. But I really want EXACTLY the same setup on my Dev, staging and production servers.

(This is the reason why we are ditching magento).

So do i create separate repo for my plugins, then write a script to pull the reaction commerce repo to my deployment pipeline, and pull the plugin repo and combine them then run and test it. then write a script to get EB to pull the reaction repo and my plugin repo and combine them to run ?

It would be ideal to make a new meteor project that depended on reaction commerce and on my plugins then the build pipeline grabs the repo installs the dependencies, tests it, packages it and send it to EB. That would be normal for me. But they seem not to do it.

Or I can Fork reaction commerce to add my plugins, but that seems wrong.

Or can someone tell me which gitignore files I need to exclude, so that I can install other dependencies later and not have a massive repo.

Someone please help.