keepalive in padrino and heroku easy xd

readme.md

Heroku Keepalive is a little ruby script that uses the Clockwork gem to keep your free heroku apps up and responsive. It does so by simply pinging whatever endpoints you specify at a specific interval.

Test locally (if you want)

foreman start

12:32:32 keepalive.1  | started with pid 9702
12:32:33 keepalive.1  | I, [2012-07-06T12:32:33.386622 #9702]  INFO -- : Starting clock for 1 events: [ ping ]
12:32:33 keepalive.1  | I, [2012-07-06T12:32:33.386750 #9702]  INFO -- : Triggering 'ping'
12:32:33 keepalive.1  | Contacted url1 in 329ms
12:32:33 keepalive.1  | Contacted url2 in 260ms
12:32:34 keepalive.1  | FAILED to contact url3 -- getaddrinfo: nodename nor servname provided, or not known

Start the keepalive process with heroku

heroku scale keepalive=1

Gemfile

source 'http://rubygems.org'
ruby "1.9.3"

gem 'httparty'
gem 'clockwork'
gem 'foreman'

keepalive.rb

require 'rubygems'
require 'clockwork'
require 'httparty'

include Clockwork
@urls = ["url1","url2","url3"]

if ENV['INTERVAL']
  interval = ENV['INTERVAL'].to_i
  abort "INTERVAL must be value in seconds greater than 0" unless interval > 0
else
  interval = 10
end

abort "ENDPOINTS must be comma-separated list of http endpoints to query for keepalive" unless ENV['ENDPOINTS']

every(interval, 'ping') {
  @urls.each do |endpoint|
    s = Time.now
    begin
      HTTParty.get(endpoint)
      duration = (Time.now - s).to_f * 1000
      puts "Contacted #{endpoint} in #{duration.round}ms"
    rescue Exception => e
      puts "FAILED to contact #{endpoint} -- #{e.message}"
      #puts e.backtrace.collect{|line| "    " + line}.join("\n")
    end
  end
}

Procfile

keepalive: foreman run clockwork keepalive.rb