I was expecting max-total-instances to control the number of app containers, but it seems to apply to the number of seats instead. This doesn’t make sense to me so I guess I’m misunderstanding something.
I’m running shinyproxy in kubernetes and have the following app-level configuration:
seats-per-container: 5
minimum-seats-available: 1
max-total-instances: 1
allow-container-re-use: true
when I launch the app in browser 1, I see that the proxy is started and a seat is consumed. When I open new tabs in the same browser to the app, I expected mores seats to be consumed–but this is not the case. So even though these new tabs are using separate sessions of the shiny app, they aren’t consuming seats. Is this by design? I thought seats were more or less equivalent to sessions?
Now when I open browser 2, and try to open the same app it gives me the following error:
“The server does not have enough capacity to start this app, please try again later.”
Which suggests to me that the max-total-instances is actually limiting seats not app instances (containers)? What am I missing here?
First, I expect sessions to be equivalent to seats. And second, I expect max-total-instances to error out only when I’ve requested more sessions (seats) than seats-per-container would allow. Would greatly appreciate some insight here.