The current logic will take the first free Process Group ID. The intention was that the process group ID space was used efficiently. The issue with that approach is that if a process group fails to come up, we can see cases where a new process group gets the same name as the old one, even thought those are actually different process groups. This can be confusing when debugging issues. A better approach would be to increase the process group IDs until we hit an upper limit and then start from the beginning.
What would you like to be added/changed?
The current logic will take the first free Process Group ID. The intention was that the process group ID space was used efficiently. The issue with that approach is that if a process group fails to come up, we can see cases where a new process group gets the same name as the old one, even thought those are actually different process groups. This can be confusing when debugging issues. A better approach would be to increase the process group IDs until we hit an upper limit and then start from the beginning.