How I spent two weeks hunting a memory leak in Ruby
Reduce Rails Memory Usage, Fix Leaks, R14, Save Money on Heroku [4 Tips]
Dynoscale | Why is Queue Time more important than Backend Request Time to auto scaling?
Why to use PUMA in production for your Rails App – Josh Software
Ruby Language Metrics (Public Beta) | Heroku Dev Center
Why to use PUMA in production for your Rails App – Josh Software
Pumaサーバーを導入して、RailsアプリをHerokuへデプロイする。 - Qiita
How many Heroku dynos do you need, and which size—An opinionated guide - Rails Autoscale
ruby on rails - Heroku, Puma, Redis, Sidekiq and Connection Limits , ERR max number of clients reached - Stack Overflow
Setup Ruby on Rails with Sidekiq on Heroku | Ruby on Fire
Why to use PUMA in production for your Rails App – Josh Software
Pumaサーバーを導入して、RailsアプリをHerokuへデプロイする。 - Qiita
Why to use PUMA in production for your Rails App – Josh Software
Why to use PUMA in production for your Rails App – Josh Software
Dynoscale | Why is Queue Time more important than Backend Request Time to auto scaling?
Vladimir Dementyev on Twitter: "@heroku Ok, my Heroku app is back :) The last but not least candidate — jemalloc. P.S. Broadcasting performance with jemalloc is only ~5-10% better than "standard" MRI.