No, `max_conns` is the maximum number of file descriptors that can be open where as `max-persistent-conns` is the maximum number of persistent connections you're referring to. Depending on your configuration, you could have a dedicated web socket server balanced by an HAPROXY server. This would allow for much greater number of connections while all serving at the same endpoint.

With concurrency in mind, `max_conn` could probably be left at the default of 1024. However, the default for `max_persistent_conns` is 100.

Have a look at this branch compare. (https://github.com/driftingrub...

You basically will need to temporarily create a file to store it and pass the path to ActiveJob. While this is currently working on local filestore, it can also work with S3 or fog.

The * is a splat operator. So, take the following example.

[2] pry(main)> fields
=> ["first_name", "last_name"]
[3] pry(main)> User.first.attributes.values_at(fields)
User Load (0.5ms)SELECT`users`.* FROM `users` ORDER BY `users`.`id` ASC LIMIT 1
=> [nil]
[4] pry(main)> User.first.attributes.values_at(*fields)
User Load (0.5ms)SELECT`users`.* FROM `users` ORDER BY `users`.`id` ASC LIMIT 1
=> ["System", "Administrator"]

http://blog.honeybadger.io/rub...

The ! at the end of find_or_create_by will raise an exception on validation errors.

Keep in mind that if you are using ActionCable on production, it does not currently work with Apache + Passenger. Instead use Nginx + Passenger.

https://www.phusionpassenger.com/library/config/apache/action_cable_integration/