Monitoring and debugging

[ncclementi]

We have something written here https://github.com/NVIDIA/accelerated-computing-hub/blob/main/tutorials/gpu-deployment/gpu-deployment-from-scratch.md#monitoring-and-debugging-tools

But we need to update this section

• Introduction to monitoring tools
  • What does each tool brings to the table
• Replace pynvml section for cuda.core
• Using the tools in a bottle neck scenario
  • Xarray example ( a starting point is in [WIP] Debbuguing and performance guide  #687)
  • Nsight / nsys
  • create cProfile and visualize with snakeviz