Background

This document explains how to create a health-check page that HAProxy will use to check the status of a web server or application.

While the exact contents of your health-check page are not important, its name should be unique (e.g. including a random number). The same page is used for all application servers to determine whether the server is running (up). While you are not prevented from using index.html as the health-check page, this is not recommended. This is because most web sites have an index.html page and, thus, there is a risk that your load balancer will direct client traffic to a web site other than your own in the cloud. For this reason, you should always use a health-check URI with a unique file name.

Example: The Amazon EC2 cloud recycles IP addresses; so, if one server was terminated and another launched for a different site—with the same IP address and page name as your health-check page name (specified in your HEALTH_CHECK_URI RightScript input)—HAProxy could consider the server to be running and part of the load-balancer pool even though it is someone else's, and direct traffic to it.

Answer

Follow the steps below to set up a Health Check Page for HAProxy for your application (Tomcat, Rails, PHP).

When following the below examples, substitute occurrences of <######> or ###### with your own unique, random numeric string.

Warning: When specifying a HEALTH_CHECK_URI RightScript input, remember to include the health-check file in your application bundle. For example, in the case of Apache Tomcat, if you set HEALTH_CHECK_URI to /health\_check378923.jsp, remember to include health_check378923.jsp in your WAR file.

Set up a Health Check Page

Apache Tomcat

  1. Create a file named health_check<######>.jsp in your application with the following contents. 

    <HTML>
  <HEAD>
    <TITLE>Hello World</TITLE>
  </HEAD>
  <BODY>
    <H1>Hello World</H1>
    Today is: <%= new java.util.Date().toString() %>
  </BODY>
</HTML>

  2. Set your HEALTH_CHECK_URI script input to:

    /health\_check_<######>_.jsp

Ruby on Rails/Mongrel

  1. Create a file named mongrel_health_check.rb with the below contents, and add it to your server or repository. 

    ####
# Copyright (c) 2011 RightScale, Inc, All Rights Reserved Worldwide.
#
# THIS PROGRAM IS CONFIDENTIAL AND PROPRIETARY TO RIGHTSCALE
# AND CONSTITUTES A VALUABLE TRADE SECRET. Any unauthorized use,
# reproduction, modification, or disclosure of this program is
# strictly prohibited. Any use of this program by an authorized
# licensee is strictly subject to the terms and conditions,
# including confidentiality obligations, set forth in the applicable
# License Agreement between RightScale.com, Inc. and
# the licensee.
#
# Handler for status/health checks.
# Load balancers (or other machines for that matter) will be able to monitor the health of
# each mongrel by retrieving a successful response from this handler
# This file can be included in the configuration of the mongrels (i.e., mongrel_cluster.yml)
# config_script: lib/mongrel_health_check_handler.rb

# This must be called from a Mongrel configuration...
class MongrelHealthCheckHandler < Mongrel::HttpHandler
  def initialize
    #Make sure it's expired by the time we process the first request
    @DB_OK_at= Time.at(0)
    @freshness= 30
    @error_msg=""
  end
  def process(request,response)

    # Write down if it's time to do a more heavyweight check
    db_stale=true if( (Time.now()-@DB_OK_at).to_i > @freshness )

    check_db if db_stale

    code = ( (Time.now()-@DB_OK_at).to_i > @freshness )? 500:200
    # Return OK if ActiveRecord is not connected yet (i.e., test mongrel only)
    code = 200 unless ActiveRecord::Base.connected?

    response.start(code) do |head,out|
      head["Content-Type"] = "text/html"

      t = Time.now()
      out.write "Now: #{t} , DB OK #{(t-@DB_OK_at).to_i}s ago\n"
      out.write "ERROR:#{@error_msg}" if @error_msg != ""
    end
  end

  # Check health of DB, and update the @DB_OK_at timestamp if it succeeds
  def check_db
    if ActiveRecord::Base.connected?
      begin      
        ActiveRecord::Base.connection.verify!(0) #verify now (and reconnect if necessary)
        ActiveRecord::Base.connection.select_value("SELECT NOW()")
        @DB_OK_at = Time.now
        @error_msg = ""
      rescue Exception => e
        # Do your logging/error handling here
        @error_msg = e.inspect
      end
    end
  end
end

uri "/mongrel-status######", :handler => MongrelHealthCheckHandler.new, :in_front => true

  2. Modify the last line in the above file so that the page can be uniquely identified with your application (e.g. change mongrel-status_######_ to mongrel-status1234089).

  3. Add the following to /etc/mongrel_cluster/mongrel_cluster.yml:
    config\_script:&nbsp; _<PathToFile>_/mongrel\_health\_check.rb Or, alternatively, set your OPT_MONGREL_CONFIG_SCRIPT input to the path where you saved the above file. Be sure to specify the full path to the file from the base directory of your Rails application; e.g. lib/mongrel\_health\_check\_handler.rb.

  4. Set HEALTH_CHECK_URI to the value you specified at the end of the file above: /mongrel-status_<######>_

PHP

  1. Create a file named health_check_<######>_.php in your application with the following contents. 

    <?php
phpinfo();
?>

  2. Set  OPT_HEALTH_CHECK_URI to: /health_check_<######>_.php

Verify Health Check Pages

If your site's permissions allow it (and assuming that your script input OPT_LB_STATS_URI is set to haproxy-status) you can simply add /haproxy-status to your load-balancer hostname URL to display an HAProxy status report in a tabular HTML format:

Example: http://www.mysite.com/haproxy-status

faq-HAProxyStatus.png