[转]How to speed up Magento 2. Maximum Performance

本文转自：https://magedirect.co/how-to-speed-up-magento-2-maximum-performance/

Introduction

In this article, we’ll show you how to run a Magento 2 store at maximum speed. We will review the most popular techniques that you can quickly implement today.

So, let’s imagine the situation:

You are preparing for big sales and the development team is making final edits in the code.
You are not happy with the current performance of your store on Magento 2 and want to finally get it over with, and to speed up it.
You have many products and categories, and the store is slow

If you find your situation with this line, then read on.

1. Preliminary preparation of the test store

Let’s create a real situation, and simulate a large store with a huge amount of products.

Thanks to the Magento 2 developers, they took care of the fact that we were able to generate a set of test products. To do this, Magento 2 has a data generator that allows you to generate a large amount of data (products, websites, categories, customers, order, etc)

More detailed instructions can be found in the official documentation:

http://devdocs.magento.com/guides/v2.2/config-guide/cli/config-cli-subcommands-perf-data.html

So we begin the fun:

Install the fresh Magento 2 on the server and load the test data into the Magento.

In this example, we will use one server for the application and the database.

We use dedicated server – DELL PowerEdge ™ R730 DX151 with the following characteristics:

CPU(s): 16

Model name: Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz

CPU MHz: 1200.281

L3 cache: 20480K

Memory: 64 GB

For speeding up the highload store, we will install the following software which is recommended by Magento Team:

Mysql
Nginx
php7-fpm
Redis
Varnish

Next, we will configure Cache Settings:

We will use the Varnish as Caching Application

Let’s enable Varnish in Magento 2 admin panel

Stores -> Configuration -> Advanced -> System -> Full Page Cache and select next settings:

Then we will enable Redis as default cache (default array) and the full page cache.

For default settings add next records in your app/etc/env.php

PHP

array ( 'frontend' => array ( 'default' => array ( 'backend' => 'Cm_Cache_Backend_Redis', 'backend_options' => array ( 'server' => '127.0.0.1', 'port' => '6379', 'persistent' => '', 'database' => '0', 'force_standalone' => '0', 'connect_retries' => '1', 'read_timeout' => '10', 'automatic_cleaning_factor' => '0', 'compress_data' => '1', 'compress_tags' => '1', 'compress_threshold' => '20480', 'compression_lib' => 'gzip', ), ), 'page_cache' => array ( 'backend' => 'Cm_Cache_Backend_Redis', 'backend_options' => array ( 'server' => '127.0.0.1', 'port' => '6379', 'persistent' => '', 'database' => '1', 'force_standalone' => '0', 'connect_retries' => '1', 'read_timeout' => '10', 'automatic_cleaning_factor' => '0', 'compress_data' => '0', 'compress_tags' => '1', 'compress_threshold' => '20480', 'compression_lib' => 'gzip', ), ), ), ),

array (

'frontend' =>

array (

'default' =>

array (

'backend' => 'Cm_Cache_Backend_Redis',

'backend_options' =>

array (

'server' => '127.0.0.1',

'port' => '6379',

'persistent' => '',

'database' => '0',

'force_standalone' => '0',

'connect_retries' => '1',

'read_timeout' => '10',

'automatic_cleaning_factor' => '0',

'compress_data' => '1',

'compress_tags' => '1',

'compress_threshold' => '20480',

'compression_lib' => 'gzip',

'page_cache' =>

array (

'backend' => 'Cm_Cache_Backend_Redis',

'backend_options' =>

array (

'server' => '127.0.0.1',

'port' => '6379',

'persistent' => '',

'database' => '1',

'force_standalone' => '0',

'connect_retries' => '1',

'read_timeout' => '10',

'automatic_cleaning_factor' => '0',

'compress_data' => '0',

'compress_tags' => '1',

'compress_threshold' => '20480',

'compression_lib' => 'gzip',

2. Generate test data for our store

We will use generated data from Magento 2. More details you can find here:

http://devdocs.magento.com/guides/v2.2/config-guide/cli/config-cli-subcommands-perf-data.html

We use so-called “Large profile”

It is:

Name	Value
website	5
store group	5
store views	5
simple products	300,000
configurable_products	8,000 with 24 options
categories	3000
categories_nesting_level	5
catalog_price_rules	20
catalog_target_rules	5
cart_price_rules	20
cart_price_rules_floor	2
customers	5000
tax rates	40,000
orders	100,000

Start the generation with the command:

Shell

bin/magento setup:perf:generate-fixtures /var/www/html/magento2/setup/performance-toolkit/profiles/ce/large.xml

1	bin/magento setup:perf:generate-fixtures /var/www/html/magento2/setup/performance-toolkit/profiles/ce/large.xml

The execution of this script takes time. If you do a similar job, we advise you to set enough memory for PHP or only personally to the given script with the help of the directive “memory_limit”. For example:

Shell

php -d="memory_limit=-1" setup:perf:generate-fixtures /var/www/html/magento2/setup/performance-toolkit/profiles/ce/large.xml

1	php -d="memory_limit=-1" setup:perf:generate-fixtures /var/www/html/magento2/setup/performance-toolkit/profiles/ce/large.xml

In this example, we set up unlimited memory for php processes.

After the script is executed in our store, we load all the test data we need to check the performance. Let’s open some test category.

Cool! The store is filled with test data. Check the administrative panel.

Products:

Orders:

A lot of categories:

Customers:

3. Let’s setup Magento for high-load

Now we will show you the most priority and simplest actions to speed up your store and to turn into a high-performance mode.

Step 1. Enable production mode
Let’s enable Magento 2 in the combat mode so that your Magento works in high-efficiency mode. For this, you need to enable production mode.

Magento 2 has different modes:

default
developer
production
More details about modes you can check here: http://devdocs.magento.com/guides/v2.0/config-guide/bootstrap/magento-modes.html

Run the following command:

Shell

magento deploy:mode:set production

1	magento deploy:mode:set production

Check:

Shell

magento deploy:mode:show

1	magento deploy:mode:show

You should see the next output:

Current application mode: production. (Note: Environment variables may override this value.)

Then run the command

Shell

php bin/magento setup:di:compile

1	php bin/magento setup:di:compile

This will then re-compile the files. Lastly:

Shell

php bin/magento setup:static-content:deploy

1	php bin/magento setup:static-content:deploy

If Magento has enabled maintenance mode – disable maintenance mode by the command:

Shell

./bin/magento maintenance:disable

1	./bin/magento maintenance:disable

Step 2. Enable Cache

We enable all cache types in the administrative panel in the section System -> Cache Management

Step 3. Flat Catalog

Let’s Enable Flat Catalog. With this setting, Magento 2 will make less complex MySQL queries in the database.

Don’t forget to clear configuration cache after this action.

Step 4. Merge & Minify CSS and JS Files

Further, we include the minification and Merge in Store -> Configuration -> Developer

Enable Merge CSS and JS Files, Minify CSS & JS. It will speed up the loading of CSS and JS

Step 5. Tune Server Configuration

Let’s show the most needed configurations for a store with a high load:

Mysql (config file: my.cnf)

The max_allowed_packet directive is chosen empirically to avoid the error “General error: 2006 MySQL server has gone away”. By studying the MySQL logs, you will be able to determine what size of packages your database needs.

For this amount of data we will use:

PHP

max_allowed_packet = 32M

1	max_allowed_packet = 32M

The directives tmp_table_size and max_heap_table_size are chosen experimentally. With the default value of Magento, many temporary tables will be created. It is defined using the command:

Shell

show global status LIKE 'Created_tmp_disk_tables';

1	show global status LIKE 'Created_tmp_disk_tables';

For this demo store we use:

Shell

tmp_table_size = 128M max_heap_table_size = 128M

1 2	tmp_table_size = 128M max_heap_table_size = 128M

The innodb_buffer_pool_size directive is set relative to the amount of RAM. The recommended value is 70-80%.

For our server this value is:

Shell

innodb_buffer_pool_size = 48G

1	innodb_buffer_pool_size = 48G

PHP (config file .php.ini)

The RAM limit must be at least 1G

Shell

memory_limit=1G

1	memory_limit=1G

For cron jobs we personally set at least 4GB

Shell

* * * * * /usr/bin/php -d=”memory_limit=”4000M” /var/www/magento2/bin/magento cron:run | grep -v "Ran jobs by schedule" >> /var/www/magento2/var/log/magento.cron.log * * * * * /usr/bin/php -d=”memory_limit=”4000M /var/www/magento2/update/cron.php >> /var/www/magento2/var/log/update.cron.log * * * * * /usr/bin/php -d=”memory_limit=”4000M /var/www/magento2/bin/magento setup:cron:run >> /var/www/magento2/var/log/setup.cron.log

* * * * * /usr/bin/php -d=”memory_limit=”4000M” /var/www/magento2/bin/magento cron:run | grep -v "Ran jobs by schedule" >> /var/www/magento2/var/log/magento.cron.log

* * * * * /usr/bin/php -d=”memory_limit=”4000M /var/www/magento2/update/cron.php >> /var/www/magento2/var/log/update.cron.log

* * * * * /usr/bin/php -d=”memory_limit=”4000M /var/www/magento2/bin/magento setup:cron:run >> /var/www/magento2/var/log/setup.cron.log

Note:

– Reindex requires more RAM on so many products, if 4GB is not enough for you – try using the bigger value.

– reindex on so many products is performed for a long time, so it’s better to set the reindex mode “update on save”

Step 6. Enable Gzip Compression

Turn on gzip compression in the nginx config:

Shell

gzip on; gzip_disable "msie6"; gzip_comp_level 6; gzip_min_length 1100; gzip_buffers 16 8k; gzip_proxied any; gzip_types text/plain text/css text/js text/xml text/javascript application/javascript application/x-javascript application/json application/xml application/xml+rss image/svg+xml; gzip_vary on;

gzip on;

gzip_disable "msie6";

gzip_comp_level 6;

gzip_min_length 1100;

gzip_buffers 16 8k;

gzip_proxied any;

gzip_types

text/plain

text/css

text/js

text/xml

text/javascript

application/javascript

application/x-javascript

application/json

application/xml

application/xml+rss

image/svg+xml;

gzip_vary on;

Step 7. Optimize Images.

There are different strategies for optimizing images. You can use third-party extensions or optimize before upload images to Magento. In this article, we use a simple script that uses open-source Linux libraries.

Just install next software:

Shell

apt-get install optipng apt-get install jpegoptim apt-get install gifsicle

apt-get install optipng

apt-get install jpegoptim

apt-get install gifsicle

Shell

#!/bin/bash find ./media -iname '*.gif' -exec sh -c 'gifsicle -b -O3 -o "{}.out" "{}"; mv "{}.out" "{}"' \; find ./media -iname '*.png' -exec optipng -o5 -keep -preserve '{}' \; find ./media -type f -iname '*jpg' -exec sh -c 'jpegtran -outfile "{}.out" -optimize "{}"; mv "{}.out" "{}"' \;

#!/bin/bash

find ./media -iname '*.gif' -exec sh -c 'gifsicle -b -O3 -o "{}.out" "{}"; mv "{}.out" "{}"' \;

find ./media -iname '*.png' -exec optipng -o5 -keep -preserve '{}' \;

find ./media -type f -iname '*jpg' -exec sh -c 'jpegtran -outfile "{}.out" -optimize "{}"; mv "{}.out" "{}"' \;

4. Let’s make final performance tests

Preliminary information:

1) We made a basic tuning Magento 2. It can be done even better personally under the store.

2) The pages for which we do Magento site speed test, we open in the browser before. That would get the page in the Varnish cache, it said “warmed up.” Because on this demo site we do not have a module for cache warmer and there are no more users who “warm up” the site.

3) In this article, we will not discuss the optimization of render-blocking JavaScript and CSS and JS minifications.

We think that a good mark for the store is 70+. A real store can have many different external scripts (chat rooms, analytics, marketing tools, etc). They all influence Google PageSpeed Insight.

Testing Magento speed (site load time) is necessary. Speed is an important metric in SEO, and also affects the user experience.
When a Googlebot indexes a site, it does not see the PageSpeed indicator, but only knows the speed itself.

Did you know that Google PageSpeed Insights does not measure the speed of your site? Read more detailed:

Google PageSpeed Insights does not measure the speed of the site.

To measure the download speed, you can use GTMetrix or Pingdom Tools.

From our point of view, there is no point in scrolling for Google PageSpeed indicators. The main speed of the store, as this affects the behavioral factor, it is very important for SEO.

Let’s look at the results of Google PageSpeed

1) Google PageSpeed Insight

Category page:

Product Page:

Configurable Product

2) Pingdom

Location: Stockholm, Sweden

Speed of category page:

Speed of product page (configurable product):

5. Additional steps that you can use for performance optimization.

CDN
Cache Warmer
Nginx professional tuning
Docker Container & Kubernetes
image optimization

All these methods we will consider in our next articles and we will reveal each method of Magento Speed Optimization in more details.

Thank you for your attention. Hope that now you understand how to speed up Magento site. Keep up with our new performance workshops!

Valentyn KubrakChief Operating Officer

Visit Valentyn's Profile Read more posts by Valentyn