Nsys profile doesn't return detail information of cuda/nvtx from docker container

ran command:
nsys profile --trace=cuda,nvtx -y 20 -d 60 -o nsyslog/nv_docker_mnist_3
docker run --rm -it --gpus all --ipc=host nvcr.io/nvidia/tensorflow:21.11-tf2-py3 python mnist.py

the nsys log shows no detail of cuda kernel details.

so I run ‘nsys status -e’ inside docker container. it returns:

Sampling trigger event available: Fail
Intel(c) Last Branch Record support: Not Available
Sampling Environment: Fail

does nsys run in docker container correctly?

thanks,

Moved to the Nsight Systems category.

Did you find out?

You might need to add --privileged=true switch and --cap-add=SYS_ADMIN switch to your docker run command.
See here:
https://docs.nvidia.com/nsight-systems/UserGuide/index.html#docker-profiling