I am deploying an R-shiny based application which includes Plotly-based interactive plots. These illustrates up to ~10000 data points across four scatter plots. The user can interact with these and see how the selected points are distributed across the three other plots. This is running well on my local computer (Ubuntu 18, 32GB RAM).
I am deploying my application to a local server (Ubuntu 16, 100+GB RAM). Everything works also on the server, but generating the plots which takes maximum a few seconds locally can take up to a few minutes on the server, making it effectively unusable.
I also run RStudio server on the same server. I have tried running the same application through RStudio Server (on the same computer that is). The data upload is very slow here, but illustration itself is as rapid as on my local computer, indicating to me that it isn't a hardware issue.
I have inspected the profiling locally and see that there are parts I could and will optimize, but this does not explain the vast difference when running on the server. I am not sure how to run the profiler for the application in deployment on the server, so I have not yet profiled it there.
Do anyone have any ideas on how to approach this? If needed I would be happy to put together and share a minimal example simulating the problem.