We have experienced some issues when deploying custom code to our ElastiCubes with custom Python algorithms. We have developed some algorithms in Jupyter Notebooks in local boxes (using Anaconda) and the execution performance of the algorithm it's about minutes.
When we deployed these algorithms in Sisense takes hours to complete. Do you have any idea about what are the pods, components that are affected by the execution of custom code in Sisense? Seems that it's a resouces trouble. Server has many more resources that local laptops, but it spent much more time to execute the same notebooks.
Thanks in advance!
If you were testing out Notebooks as part of a beta release, it is likely you ran into some known limitations of the beta with regard to executing Python.
Additionally, there are in fact some performance considerations as Notebooks makes use of a kernel (virtual machine) that may take a couple of minutes to start once the first code block is executed. You can find those details here.
If you are now on the latest release and still running into performance issues, please feel free to check in with our Support Team who can help you troubleshoot. You can open a ticket at: support.sisense.com
If that is the case, Support can assist you with troubleshooting and running diagnostics to help find out where the issue lies. Feel free to open a ticket at https://support.sisense.com
All the best,