PHPDay 2013 - High Performance PHP

High Performance PHPPHPDay 2013Jonathan Klein@jonathanklein

Saturday, May 18, 13

Slides, Links:jkle.in/phpday

Some Etsy Stats• 1.4 billion page views/month• Almost $1B in sales last year• Over 1M lines of PHP

Agenda• Why Performance Matters

• Profiling PHP Applications

• Code Level Optimizations

• Big Wins

• Load Testing

• Takeaways

The Value of Performance

Real World Examples

http://www.phpied.com/the-performance-business-pitch/

Real World Examples• Firefox: -2.2 seconds = 15.4% more downloads

• Shopzilla: -5 seconds = 7-12% increase in revenue

• Google: +400ms = 0.76% fewer searches

• Amazon: +100ms = -1% revenue

~80% of page load time takes

place on the client

...if your backend is fast

A fast page load is 2 seconds

This means you have 400ms to get that HTML off your server

But network time could be ~100ms

This means you have 400ms 300ms to build the page

< 100ms feels instant< 1 sec feels like flow

< 10 sec to keep user’s attention

http://www.nngroup.com/articles/response-times-3-important-limits/

Full Page Load – 2 Seconds

Base HTML – 400ms

Server Generation Time – 300ms

Profiling PHP Applications

Monitoring/Tracing • Paid:

• Tracelytics (bought by AppNeta)

• AppDynamics (building a PHP solution)

• dynaTrace (building a PHP solution)

• New Relic (has a free option)

• Free:

• StatsD/Graphite

• xhprof

Monitoring/Tracing • Paid:

• Tracelytics (bought by AppNeta)

• AppDynamics (building a PHP solution)

• dynaTrace (building a PHP solution)

• New Relic

• Free:

• StatsD/Graphite

• xhprof

StatsD (UDP packets)$start = microtime(true);/* script content */$end = microtime(true);

StatsD::timing('foo.bar', $end - $start);

More Info: http://goo.gl/LbDPE

Graphite• Written by Orbitz

• Real-time graphing engine for StatsD data (among other things)

• http://graphite.wikidot.com/

• Architecture: http://www.aosabook.org/en/graphite.html

Etsy Conversations PHP Time

Execution Count

Search Page PHP Time (95th Percentile)

Search Page Execution Count

Stacked Search Timers

xhprof• PHP Extension (need to install)

• http://pecl.php.net/package/xhprof

• Code level tracing

• Significant overhead, use in DEV only!

• Add ?xhprof=1 to URL

• Results in browser

Lesson: Profile Your Code

Code Level Optimizations

Writing Efficient PHPSet max value before loop:

$max = count($rows); for ($i = 0; $i < $max; $i++) { echo $i; }

require_once() is slow

Minimize use of define()

Yes, single quotes are slightly faster than double quotes, but...

Almost Every Micro-Optimization is Worthless

http://phpbench.com/

Writing Efficient PHPSet max value before loop:

$max = count($rows); for ($i = 0; $i < $max; $i++) { echo $i; }

Okay, this one is pretty good

So Why Even Mention Micro-Optimizations?

http://code.flickr.net/2009/12/02/flipping-out/

strtok() to implode()

Lesson:Focus on the Big

“Premature optimization is the root of all evil”- Donald Knuth

Big Wins

Upgrade PHP

Upgrade PHP5.3 is ~20% faster than 5.2

http://news.php.net/php.internals/36484

5.4 is ~20-40% faster than 5.3

5.4 is ~20-40% faster than 5.3http://news.php.net/php.internals/57760

Upgrading 5.2 --> 5.4 gives a 45-70% improvement!

Upgrading 5.2 --> 5.4 gives a 45-70% improvement!http://php.net/migration53

http://php.net/migration54

PHP 5.4 is 5 Times Faster than PHP 4http://static.zend.com/topics/White-paper-PHP4-PHP5.pdf

Etsy’s Upgrade to PHP 5.4

Use an Opcode Cache (APC)

Standard Page Execution

With An Opcode Cache

Opcode Cache Vanilla settings: 30-40% improvement

Turn off APC Stat: additional ~2x improvement -- Understand what is happening herehttp://www.slideshare.net/vortexau/improving-php-application-performance-with-apc-presentation

Cache Data in a Key-Value Store

APC User Cache<?php$foo = "Hello, World!";apc_store('some_key', $foo);?>

<?phpvar_dump(apc_fetch('some_key'));?>

-------- Output --------

string(12) "Hello World!"

APC User Cache

APC User Cache • Avoid fragmentation - keep utilization under 10%

• Assign 1GB, only fill 100MB

• Compress objects that are > 10KB before storing

• Reduce garbage collection in the source of apc_store()

• Consider CDB

• Consider CDB• http://engineering.wayfair.com/moving-constants-out-of-apc-

and-into-cdb/

Memcached• Usually a separate server• In-memory key-value store• Extremely simple and fast• http://memcached.org/

APC vs. MemcachedAPC User Cache Memcached

Local to the Server Shared Network Resource

Good for small objects Large or small objects

Good for mostly read workloads

Can read and write quickly

Only one instance Can be clustered

Fix All Errors• PHP 5.3: E_ALL | E_STRICT

• PHP 5.4: E_ALL

• Can also do error_reporting(-1);

Child Processes4-6 processes per CPU core.

Beyond that just add servers.

(Test your app)

HipHop for PHP• Developed/Open Sourced by Facebook

• Now a VM + JIT compilation

• 5x improvement in throughput over PHP 5.2• http://developers.facebook.com/blog/post/2010/02/02/

hiphop-for-php--move-fast/

• https://www.facebook.com/notes/facebook-engineering/speeding-up-php-based-development-with-hiphop-vm/10151170460698920

• https://github.com/facebook/hiphop-php

Understand Framework

Overhead

http://systemsarchitect.net/performance-benchmark-of-popular-php-frameworks/

Symfony is 20x slower than raw

Load Testing

JMeter• Open Source• Generate load via a GUI or command

line• Can watch req/s degrade with more

users• Easy to use

Be Careful

Be Careful• JMeter looks a lot like a DOS attack

Be Careful• JMeter looks a lot like a DOS attack• Make sure you know what is failing

Be Careful• JMeter looks a lot like a DOS attack• Make sure you know what is failing• Look at monitoring while test is running

Be Careful• JMeter looks a lot like a DOS attack• Make sure you know what is failing• Look at monitoring while test is running• Run in Production

• Run a test, make a change, run it again

Takeaways

“How Fast Is Your Site?”

This is a terrible quesiton

Why is it terrible?

Why is it terrible?• Lack of context

• Are we talking about average or a percentile?

• Server side time or client?

• Who is measuring it?

• When is it being measured?

• Real users or synthetic?

We still have to answer it

Pick Tight SLAs

“The homepage of our site will load in <300ms at the 80th percentile, measured by sampling 10% of our real users over a 24 hour period every day at 8AM.”

Pick Tight SLAs

“The homepage of our site will load in <300ms at the 80th percentile, measured by sampling 10% of our real users over a 24 hour period every day at 8AM.”

Things to Remember

Things to Remember• Measure and monitor your application

• Focus on big wins

• Run the latest (stable) version of PHP

• Make sure you are using APC correctly

• It’s always the database (stay in this room)

• Caching is your friend

• Know what system resources you depend on

There is a lot more to talk about

Get in Touchhttp://web-performance.meetup.com/

www.etsy.com/careers

jonathan@etsy.com

@jonathanklein

PHPDay 2013 - High Performance PHP

Technology

symfony & jQuery (phpDay)

Performance Tuning in PHP

PHP Profiling/performance

Php performance-talk

Developing and Deploying High Performance PHP · PDF file1.....

20100529 phpday

Programação Orientada a Aspectos - PHPDay SERPRO Curitiba

Grab yourself an alibi - PHPDay 2013

High Performance PHP - PHPNW

Application Performance Management - PHP

Keynote del phpday 2010

PHPDAY - PHP com excelência - Só para os fortes! por...

PHP Performance: Principles and tools

Building distributed systems - PhpDay 2013

Detrás del Backend [phpDay 2015]

Php Performance On Windows