Heroku Puma Rails 4 : Getting R14 errors, how do I scale my server up?

Question

I am running a Rails 4 app on Heroku, with Puma server. I am using only one basic web dyno (no worker), with 512MB ram.

Recently, I have been facing a lot of R14 errors (Memory Quota Exceeded), my server crashed three times yesterday.

Here's my Puma config file :

workers Integer(ENV['WEB_CONCURRENCY'] || 0)
threads_count = Integer(ENV['MAX_THREADS'] || 5)
threads threads_count, threads_count

preload_app!

rackup      DefaultRackup
port        ENV['PORT']     || 3000
environment ENV['RACK_ENV'] || 'development'

on_worker_boot do
  # Worker specific setup for Rails 4.1+
  # See: https://devcenter.heroku.com/articles/deploying-rails-applications-with-the-puma-web-server#on-worker-boot
  ActiveRecord::Base.establish_connection
end

I have few time-consuming tasks being performed each day on my server, I think they are the reason why I am getting these R14 errors and server crashes :

• rake tasks being performed with Heroku Scheduler

• imports of 10 photos (size max : 10 x 3MB) when creating / updating a model named Thing

• exports of logs from my database, that can contain up to 6k rows

Here are my questions :

1. Will putting these time-consuming tasks on a new worker dyno help me avoid R14 errors / server crashes ? If yes, do you recommend using delayed_job_active_record gem or resque gem?

2. Do you think my Puma config file is ok ?

3. I can't figure out how to put rake tasks performed by Heroku Scheduler on a worker dyno. Any idea ?

Thanks a lot for your help !

Answer 1

I have no personal experience with setting up Heroku worker dynos, but as for your question #2 the file should be fine as long as you are running Rails 4.1 or newer, otherwise the config should match the recommendation in this link: https://devcenter.heroku.com/articles/deploying-rails-applications-with-the-puma-web-server#on-worker-boot.

In a recent attempt to upgrade my app to Rails 4.2, memory consumption raised significantly and R14 errors became much more frequent. So I downgraded back to Rails 4.1 (Ruby version is 2.2.2). Maybe that's something you could give a try.

Also, if you haven't done so, I recommend you installing the New Relic add-on and taking a look at the per-instance memory consumption at the "Monitoring/Instances" link.