LGs must have sufficient upstream bandwidth if you’re sending large requests or testing streaming protocols. Also, ensure low latency between LGs and the target environment if possible.