Julik Tarkhanov
Julik Tarkhanov
👍 Thanks, we will do some stress testing.
We did. We ran falcon with 6 worker processes, putting it behind nginx on one of our production instances (where puma used to run instead). We had to switch to...
We are using HTTP/1.1 from falcon to nginx, and HTTP/1.1 from nginx to CloudFront which is our fronting CDN. HTTP/2 is not in the picture for us at the moment....
The ideal would be that if bytes per second for a client is below N bytes per second over N seconds I would kick the client out. However, I can...
Yep, being able to set a timeout for each `read` and each `write` would be ideal. What I effectively have attempted with my polling solution is actually doing the `write`...
Having investigated a bit, would this work? Specifically, will it "override" a Condition? ```ruby TimeoutOnWrite = Struct.new(:async_task, :writable, :timeout_s) do def write(data) async_task.timeout(timeout_s) { writable.write(data) } end end body =...
Yep, tried that implementation and though the timeout does fire it brings down the reactor (and the entire falcon process as well!) I will revert to my less-than-ideal polling implementation...
Happy 2019 @ioquatix and other socketry contributors! We have deployed our falcon-based service as a canary and observing the results. Meanwhile I am trying to figure out where the limits...
> Without digging into it too much (dude I'm on holiday at the beach) Man I envy you we are freezing here in the northern hemisphere 🥶 Enjoy your holidays...