No, it’s not. Relevance runs on each computer, and no client is aware of what any other client is doing.
The “free” way to handle a case like this is to use the “stagger start time” option when issuing the action. Each client will wait a random amount of time before starting the action. This doesn’t guarantee any particular number of machines running at once, but it does randomly spread the action over time.
For more precise timing you could look into the Server Automation capabilities, but I believe your runbook would need to specify which two machines to run for each batch.
Otherwise you’d need to write your own script to hand the actioning, sending the action to two computers at a time and waiting for them to post results before moving on to the next.