Dear Community,
Suppose I have a python-derived Docker image, set up to simply run a python application that is set up to be an intermediate processor of data in a kafka-based ETL pipeline. “docker run” starts that python program in the foreground and it’s continually running, consuming/producing, until it’s killed. There’s no web service, no ports need exposing. Autodevops works great for building and pushing the docker image to gitlab. It also performs a number of tests without issue. But it won’t deploy. Ultimately the image isn’t quite compatible with the “helm install” part of the auto-deploy script in the auto-deploy-image. I find it a little strange that autodevops is clearly geared for deploying web services, but the impression I get from the documentation is that it is more generic than that. So does anyone have some advice? I’m sorry if this has been discussed already, I’ve made genuine efforts to find more info and this post is a sort of last (or next-to last) resort.
It looks more and more like I’ll be needing to create my own deploy, with it’s own helm chart and let my kubernetes gitlab-runner use that, outside of the AutoDevOps framework. But just in case, are there examples of autodevops deployments that are non-web that I can see and maybe follow? One of the values in the values.yml file for the default autodevops helm chart is application.tier=“web”. Are there other “tiers”?
Thanks,
Andy