Is there a way of configuring the maximum number of requests that Plumber queues up?
I'm not aware of a queuing mechanism.
Sorry, maybe I shouldn't have used the word queue. What I mean is the maximum number of connections that the server is able to handle at the same time. However, since R is single threaded it can only process one request at a time, so the other requests that arrive end up having to wait. There is an implicit queue of (TCP ?) connections, and I would like to be able to configure this setting.
I think this goes to how you are hosting your api, for example posit connect should scale it for you, whereas other services may not ?
Hosting • plumber - load balancing
My issue is not related to being able to serve more clients. I'm trying to test my service, which has a plumber API among other components, and want to test what happens when plumber starts dropping packets/requests, because it reached its maximum, and starts sending HTTP 503 responses. However it is difficult to test this kind of events since I don't really know what that limit is, and how to control it.
my best guess is its httpuv package that would throw the 503, perhaps asking the maintainers about the conditions would be fruitfull.
Issues · rstudio/httpuv