I had a look at codecov.io and it is very simple to set up. I think it could be useful to visualize our test coverage, especially for diffs on PRs. I am aware that coverage doesn’t mean, that the code actually does what it should, but no coverage means it’s certainly untested and that maybe warrants special attention then.
All that’s required is to run tests with coverage recording enabled and a little bit of scripting and calling a bash script to upload to codecov. That can either be a separate build step/job on the CI or adding the necessary options to one of the existing test stages. Both could be done as an additional command in build.go or a separate script. I tested it for unit tests, but it should even be doable on the integration tests as well.
What do you think?