You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I wouldn't be able to use this because we configure our unicorn and resque process pools to run on a given class of hardware, although I suppose we could detect the dyno class and set parallelism dynamically on boot.
I don't have a clear vision of how the logic (and config) for this would interact with the simple "N users per dyno" or "M rpms per dyno" we use now. I guess we would need to define a 'users per dyno' value for each dyno class, but then defining the boundary of when you scale up to the next class vs adding more 1X dynos would be hard to fathom.
I think what I'm saying is that this is too complicated and doesn't answer a real need for us at Harry's, but I could see it being really useful for a site with very low average traffic that is highly burstable. A pull request would definitely be considered.
Eg in low periods scale down to hobby and scale up when needed
The text was updated successfully, but these errors were encountered: