View metrics with jupyterlab » History » Version 11

« Previous - Version 11/12 (diff) - Next » - Current version
Pengfei Ding, 10/14/2019 03:56 PM

Getting setup with docker and jupyterlab

Assuming you have docker ready, the following commands will pull a script for starting a docker container (by default named as "metrics_nb"). In the container, an instance of jupyter notebook will be started automatically. The URL to the jupyter notebook instance will be displayed in your terminal.

chmod +x

Once you get the container start script, you can pull down the notebook used for plotting metrics by:

mkdir -p notebooks; cd notebooks
cd ..

Then run the container start script it with:

# ./ <metrics_dir_path> <notebook_dir_path> [port_number] [notebook_name]
# e.g.
./ /home/ron/work/artdaqPrj/demo2-develop/daqlogs/metrics ./notebooks 8880 test_nb

Once the container is up, you can open a browser to the link displayed in your terminal. Open and run the "notebooks/Plot_metrics-single_plot.ipynb" notebook. There will be two drop-down boxes asking you which log file you want to parse and which metrics you want to plot.

You can test run this on as user "sbnd", "root" or add yourself to the "docker" group first. Remember to shutdown the container with (this message also gets printed out when starting up the container):

docker rm -f [notebook_name] #default is metrics_nb

Note: if you want to run multiple containers, or if multiple users are running the same containers, you may need to specify *port_number" and "notebook_name" to be different from the default ones to avoid conflict.