Two different ways to plug in a simple Warden failure app to a Rails project.

gistfile1.rb

##
# Simple Warden configuration, complex "controller"

# In config/initializers/warden.rb
Rails.application.config.middleware.use Warden::Manager do |manager|
  manager.failure_app = UnauthorizedController
end

# app/controllers/unauthorized_controller.rb
class UnauthorizedController < ActionController::Metal
  include ActionController::Rendering

  def self.call(env)
    @respond ||= action(:respond)
    @respond.call(env)
  end

  def respond
    render nothing: true, status: :unauthorized
  end
end

##
# Simple controller, complex Warden configuration

# config/initializers/warden.rb
Rails.application.config.middleware.use Warden::Manager do |manager|
  manager.failure_app = ->(env){ UnauthorizedController.action(:index).call(env) }
end

# app/controllers/unauthorized_controller.rb
class UnauthorizedController < ApplicationController

  def index
    render nothing: true, status: :unauthorized
  end
end