App crashes with more than 2 concurrent users without any error logging

I created an R Shiny Dashboard, and an hour ago I presented it to a group of people. The dashboard loads data from online sources (via googlespreadsheets4 and readr) and caches it on local flat files saved in CSV files using the readr package.

While the app works well when one or two people are accessing it, with about 10 people in the room all accessing it simultaneously, the app disconnects with a grey screen and a disconnection message notifying the app has been disconnected. It looks vaguely similar a timeout disconnection. When I asked everyone else in the room to stay off the app, it worked fine again and experienced no further issues.

I later checked the logs. Although there were plenty of events logs --warnings that I am used to seeing--there was no indication of any error. I wondered if readr was crashing when it was trying to read files into multiple instances simultaneously but there were no indications of any readr file read or write errors.

The app is at https://bnanalysis.shinyapps.io/safe_border_opening/.

I also checked resource usage. Memory usage does not get above 600 MB, but CPU usage did spike quite high - it got to 125000000000.0 on the graph (I don't know what the units are).

Does anyone know what might be a cause of a problem like this?

1 Like

Hello @benjsmith. Did you manage to solve this? I have a shinyapps.io account with 2 apps with each requiring at least 4 GB of RAM and probably 5 people will need to use it simultaneously.

One of my main problems in the tuning performance settings is that I do not fully understand what a worker is.