I have multiple runners and with the keyword ‘dependencies’ I get the right sequence for multiple envs builds/deploys.
My tasks do not define a specific runner and my runners can pick up untagged jobs.
The “problem” is that, if I define the build and deploy jobs with manual execution, the two jobs run in parallel and the deploy task obviously fails.
I can solve this by defining a specific tag (runner) for both jobs, but then I loose the benefit of having multiple runners.
I tried the new “needs” keyword but it does not solve my problem. What I’d like to have is a fire-and-forget solution, ie: I open the pipeline, I click on build then click on deploy and close the browser. GitLab should run the build task with a random runner and, once completed, run the deploy task with another random runner.
This is exactly the behaviour I get right now with automatic execution and I’d like to have the same with the manual execution.
Thanks, any help is really appreciated!