README.md

Pixel Tracker

This pixel tracker is built on OpenResty build of NGINX. It uses lua to generate a random UUID for each request. Lua gives you an opportunity to do some more intelligent routing if you want. All responses are logged to a file called access.log

This project serves a couple of purposes:

1. It's a simple pixel tracker that logs all requests with the raw request data to a log file
2. It's a default lua project for OpenResty that shows you how to leverage access filters and response filters.

Features:

• 1px gif response (200)
• no content gif response (204)
• ability to capture post data
• redirects

Getting Started

Assuming you are running this on OSX, you will need sudo access to run on port 80/443. This project assumes you are using the host name as the account. The best way to run this would be using a wildcard domain *.hostname.com, where * is the account (a-z0-9). If you want SSL you would need to setup a wildcard cert. The configuration file by default is setup for localhost.

For testing purposes, you will need to modify your /etc/hosts file to include some test domains. In this case I have added two domains test.localhost and test2.localhost for testing.

##
#
# localhost is used to configure the loopback interface
# when the system is booting.  Do not change this entry.
##
127.0.0.1	localhost test.localhost test2.localhost
255.255.255.255	broadcasthost
::1             localhost
fe80::1%lo0	localhost

Once created, you can start NGINX as a sudo user.

sudo nginx -p /Users/rick/Documents/workspaces/opensource/pixel_tracker/

When you request one of the application urls, you will log the response in tab separated format. Test a couple requests using your browser. To test post, use the following curl command:

curl -X POST "http://test.localhost/id_here/data" -d '{"title" : "Test", "tags" : ["foo","bar"]}' 

Here's an example of what data in the log will look like:

1389561927.133	cb9c1a5b-6910-4fb2-b457-a9c72a392d90	test	testlaptop.local	127.0.0.1	-	GET /test/uuid.json HTTP/1.1	200	test	http	test.localhost	/test/uuid.json	-	-	1	-	Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36	text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8	gzip,deflate,sdch	en-US,en;q=0.8	-	-	-	-	GET /test/uuid.json HTTP/1.1\x0D\x0AHost: test.localhost\x0D\x0AConnection: keep-alive\x0D\x0ACache-Control: max-age=0\x0D\x0AAccept: text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8\x0D\x0AUser-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36\x0D\x0ADNT: 1\x0D\x0AAccept-Encoding: gzip,deflate,sdch\x0D\x0AAccept-Language: en-US,en;q=0.8\x0D\x0A\x0D\x0A
1389562820.171	8e3cea51-20d3-440b-9aa6-a90a1bf1839c	id_here	testlaptop.local	127.0.0.1	-	POST /id_here/data HTTP/1.1	200	test	http	test.localhost	/id_here/data	-	-	-	-	curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8y zlib/1.2.5	*/*	-	-	-	-	{\x22title\x22 : \x22Test\x22, \x22tags\x22 : [\x22foo\x22,\x22bar\x22]}	-	POST /id_here/data HTTP/1.1\x0D\x0AUser-Agent: curl/7.24.0 (x86_64-apple-darwin12.0) libcurl/7.24.0 OpenSSL/0.9.8y zlib/1.2.5\x0D\x0AHost: test.localhost\x0D\x0AAccept: */*\x0D\x0AContent-Length: 42\x0D\x0AContent-Type: application/x-www-form-urlencoded\x0D\x0A\x0D\x0A	data

Other Notes

The raw response is included in your log file. NGINX encodes the response. You will see lines with \x0D\x0A etc. These are encoded for your safety. You will need to decode them to process them effectively.

To save you some time I have a couple code samples of how to do this in:

• perl
• java
• javascript

For Perl, use the following one-liner:

	$txt =~ s{\\x(([01][0-9a-f])|([2][2f])|([5][c]))}{ chr(hex($1)) }egi;

I found using the StringEscapeUtilities from the Apache Commons library does the job nicely for Java. Use the following code to make your life easier:

	protected static final String decodeEscapeSequence(String str,
			String defaultValue) {

		if (StringUtils.isBlank(str)) {
			return defaultValue;
		}
		return StringEscapeUtils.unescapeJava(
				str.replaceAll("(?i)\\\\x([0-9a-f]{2})", "\\\\u00$1")).trim();
	}

For Javascript, use the following String prototype function:

	String.prototype.decodeEscapeSequence = function() {
  	  return this.replace(/\\x([0-9A-Fa-f]{2})/g, function() {
  	      return String.fromCharCode(parseInt(arguments[1], 16));
  	  })
  	  .trim();
	};