I have just made the monthly deployment of my Shiny app. I get no error messages during the deployment, but when I access the site I get this: An error has occurred
The application failed to start.
signal: killed The app works as expected on my computer. I have no idea of how to resolve this. Do any of you?
If you can provide the url or id, we could look to see if there is something on our side. You should be able to access the application logs in the dashboard, that would be a good place to start.
Gemini put me on to the steps to access the log file. I see that I get an OOM error. The uploaded files constitute 39,365,747 bytes. I have a second app running on the free service. If I take that off-line, will that free up some more RAM for the first app?
Good to here you have made some progress here. Unfortunately, taking the second app offline will not free up memory for the first app, each application gets its own memory.
That file doesn't seem too big, maybe something the application is doing with the file is causing it to consume a lot of memory?
When I look at the log, the last lines are: 025-10-15T20:57:12.807896+00:00 shinyapps[12345495]: New names: 2025-10-15T20:57:12.812190+00:00 shinyapps[12345495]: • `` -> ...12025-10-15T20:57:12.816465+00:00 shinyapps[12345495]: New names: 2025-10-15T20:57:12.820844+00:00 shinyapps[12345495]: • `` ->...12025-10-15T20:57:13.787275+00:00 shinyapps[12345495]: New names: 2025-10-15T20:57:13.791929+00:00 shinyapps[12345495]: • `` ->...1 2025-10-15T20:57:13.797377+00:00 shinyapps[12345495]: Container event from container-12561886: oom (out of memory)
At the third instance of New names the app aborts.
When I start the app on my computer there are eight instances shown in the console before the app window appears.
I recognize this message as coming from a call to bind_rows(). I have two calls to bind_rows() in two different observe_event() instances.
In each case the bind_rows() function is followed my four mutate() calls. That makes for eight points where the New names message might appear.
This seems like where the memory issue arises. Do you have any guidance on work-arounds for this?
I set up an account, but I don't see how to publish from RStudio. I'm asked to contact my administrator for the public URL of the Posit Connect server. Where can I get the instructions I need? And, thanks.
Ahh... I forgot you can currently only deploy to posit connect cloud from public Github repos or from Positron.
You could also try using https://profvis.r-lib.org/ to record the RAM usage when you run it locally. There might be way you can refactor something to use less RAM, but it's really difficult to know without seeing your code.
After some work with Gemini, ChatGPT, and Git Hub I got the app onto Posit Cloud. It’s way more work than deploying to shiny apps.io from RStudio, but thanks for your interest. I’m hoping that Posit doesn’t bail on RStudio.