Your question is Robustness on Noisy Network Telemetry. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you evaluate the robustness of a model trained on noisy, high-velocity network telemetry data?