Hey, I started looking into my first issue and picked https://github.com/runatlantis/atlantis/issues/2685 for it. After digging into the code, I understood it's not a GitLab-related issue per se. The error occurs when Atlantis tries to set the state to running again. I reproduced the behavior in GitHub by planning a couple of projects and applying only one. This leads to the state of

atlantis/apply

job staying pending and showing the wrong number of projects. Am I correct in assuming this is a bug and the

atlantis/apply

job should transition to the success state, even when not all plans are applied? Or what should be the intended behavior here?

I think that correct behavior here is that

atlantis/apply

should transition to success state only when there are no plans to apply. Meaning that each project had it's plan applied or project was unaffected and had "No infrastructure changes" instead. And it works well this way for github from our experience.