When you assign multiple targets to a deployment step, RGDM deploys to each target in parallel.
We have a step that does not run safely in parallel. Deployments fail unpredictably because of deadlocks.
We want to run the step strictly against one target at a time to avoid the deadlocks.
Octopus Deploy supports this through the special OctopusMaxParallelism variable.
Paul Stovell explains:
you can set a special variable called OctopusMaxParallelism to a value of, say, 5. This way the step will only be executed on 5 machines at any given time. If one machine finishes, then the next will start, and so on, with only 5 ever running at once.
Does RGDM have something similar? If it existed, we would try setting it to 1.
If you don't support this yet, do you have a workaround instead?