Pipeline load balancer

We have several projects which are triggering a downstream pipeline to run integrations tests against our application. Currently we have only one application instance in use, meaning that if there are multiple projects triggering the downstream pipeline simultaneously the latter ones need to wait the earlier tests to finish. We control this by defining “resource_group” in the downstream pipeline job and it works just fine.

Now, to remove the bottleneck we are planning to set up more application instances where to run the IT tests. Lets say we will have a pool of three application instances/servers, but the question is how to configure the pipelines to handle this? Is there a “load balancer” in Gitlab that could handle this scenario?

So we would need to trigger the IT tests into one of the application instances/servers if there is one available or wait that one will get free. I guess the waiting can still be handled with “resource_group” but how to configure the load balancer part in .gitlab-ci.yaml?