Comment by fvdessen

That also works with monoliths ... Usually you would make your monolith stateless and distribute the incoming requests / events across many instances that can be spawned / killed depending on volume of requests and health status of instances.