Scaling Workers

I have modeled about ten processes, each process containing about ten service tasks (each with a worker behind), and I am wondering about scalability issues when I will have hundreds of BPMN with associated workers.
In this scenario, i would have more than a thousand workers listening. This means that my 1000 workers will be listening most of the time and could impact the performance of my machine and even crash the Zeebe broker.
What is the best approach to model/implement workers in this case?