Exactly as described, this is the per-entity ceiling on how many of this entity's configured test instances can be simultaneously enabled if it's a reflector, or running if it's a generator, constrained by available hardware sessions on that entity. Because this depends on hardware session availability rather than pure software configuration, it can be lower than the configured-instance limit for the same entity, meaning an operator can successfully configure more test instances on a port than can ever run at once. For example, an entity that can hold 8 configured test instances but only run 4 simultaneously due to FPGA session limits would require running the remaining 4 in a second wave.