You are browsing the documentation for version 3.1 of OroCommerce, OroCRM and OroPlatform, which is no longer maintained. Read version 5.1 (the latest LTS version) of the Oro documentation to get up-to-date information.
See our Release Process documentation for more information on the currently supported and upcoming releases.
Optimize PHP performance¶
PHP-FPM (FastCGI Process Manager) is an alternative PHP FastCGI implementation adjusted for better handling of the heavy workload.
The recommended configuration of the PHP-FPM is provided below.
1 [www] 2 listen = 127.0.0.1:9000 3 ; or 4 ; listen = /var/run/php5-fpm.sock 5 6 listen.allowed_clients = 127.0.0.1 7 8 pm = dynamic 9 pm.max_children = 128 10 pm.start_servers = 8 11 pm.min_spare_servers = 4 12 pm.max_spare_servers = 8 13 pm.max_requests = 512 14 15 catch_workers_output = yes
Make sure that Nginx
fastcgi_pass and PHP-FPM
listen options are aligned.
Optimize PHP Runtime Compilation¶
Use an OpCache bytecode engine to cache bytecode representation of the PHP code and save time on the repetitive runtime compilation.
Please install Opcache php-extension and configure it in the following way:
1 opcache.enable=1 2 opcache.enable_cli=0 3 opcache.memory_consumption=512 4 opcache.max_accelerated_files=65407 5 opcache.interned_strings_buffer=32 6 #http://symfony.com/doc/current/performance.html 7 realpath_cache_size=4096K 8 realpath_cache_ttl=600
The opcache.load_comments and opcache.save_comments parameters are enabled by default and should remain so for Oro application operation. Please do not disable them.
Optimize Web Server Performance¶
You can improve your website performance by turning on compression and caching. This is configured on your web server.
gzipcompression, add the following into your
nginx.confor website configuration file:
Nginx.conf is usually located at
pagespeed_modfor nginx, use the build ngx pagespeed from source guidance.
HTML compression, add the following lines into your
nginx.conf or website configuration file in sections
pagespeed on; pagespeed FileCachePath "/var/cache/ngx_pagespeed"; pagespeed EnableFilters collapse_whitespace; pagespeed Disallow "*.svg*";
Nginx.conf is usually located at
To enable caching, insert the following in the server section of your website configuration file:
If you are using Apache as your web server, you already have the necessary configuration in the
However, this configuration relies on the
mod_headers modules that are needed for the compression
and caching to work. Ensure these modules are enabled in Apache configuration.
To enable compression, ensure that
mod_deflatemodule is loaded in your Apache config file as illustrated below:
LoadModule deflate_module libexec/apache2/mod_deflate.so
Apache configuration is usually located at
The out of the box configuration for the compression in the
.htaccessfile is following:
Pagespeedmodule for Apache, follow the guidance on installing from Apache-only packages. To enable
HTML compression, ensure that these lines are uncommetned in
ModPagespeed On ModPagespeedFileCachePath "/var/cache/mod_pagespeed/" ModPagespeedEnableFilters collapse_whitespace AddOutputFilterByType MOD_PAGESPEED_OUTPUT_FILTER text/html
To enable caching, ensure that
mod_headersis loaded in your Apache config file as shown below:
LoadModule headers_module libexec/apache2/mod_headers.so
The out of the box configuration for caching in the
.htaccessfile is following:
There are a few ways to tune up search speed performance:
Give memory to the filesystem cache
Use faster drives (SSD instead of HD, local storage over virtual)
Search fewer fields
Warm up the filesystem cache
See more information on optimizing search speed on Elasticsearch website.
To tune for indexing speed, you can try the following recommendations:
Use multiple workers/threads to send data to Elasticsearch to use all resources of the cluster
Increase index.refresh_interval to allow larger segments to flush and decreases future merge pressure
Disable refresh and replicas for initial loads
Give memory to the filesystem cache
Use faster hardware
See more information on optimizing indexing speed on Elasticsearch website.
Also, keep in mind that it is not recommended to use Elasticsearch with MySQL, PostgreSQL, Redis and/or Rabbit on one server to avoid slow performance.
To optimize Redis, try the following configurations for performance optimization:
maxclients 100000 maxmemory 512mb maxmemory-policy allkeys-lru maxmemory-samples 3
Append only mode
appendonly no appendfsync everysec no-appendfsync-on-rewrite no auto-aof-rewrite-percentage 100 auto-aof-rewrite-min-size 64
slowlog-log-slower-than 10000 slowlog-max-len 1024
hash-max-ziplist-entries 512 hash-max-ziplist-value 64 list-max-ziplist-entries 512 list-max-ziplist-value 64 set-max-intset-entries 512 zset-max-ziplist-entries 128 zset-max-ziplist-value 64 activerehashing yes
The complete configuration recommendations is available in the Redis configuration file example.
You can find more information on memory optimization on Redis website.
The following recommendations that can highly improve PostgreSQL performance:
Increase the shared_buffers value in postgresql.conf. The shared_buffers parameter defines how much dedicated memory PostgreSQL uses for the cache. The recommended value is 25% of your total machine RAM, but the value can be lower or higher depending on your system configuration. Try finding the right balance by altering the values.
Increase the effective_cache_size value in postgresql.conf. The parameter specifies the amount of memory available in the OS and PostgreSQL buffer caches. Usually, it should be more than 50% of the total memory. Otherwise, it may slow down the performance.
Increase the work_mem value, if you need to do complex sorting. But keep in mind that setting this parameter globally can cause significant memory usage. So it is recommended to modify the option at the session level.
Increase the checkpoint_segments value to make checkpoints less frequent and less resource-consuming.
Increase the max_fsm_pages and max_fsm_relations value. In a busy database, set the parameter to higher than 1000.
Reduce the random_page_cost value. It encourages the query optimizer to use random access index scans.
For more optimization configurations, see PostgreSQL website.
You can get better performance and minimize storage space by using some of the techniques listed below.
Optimize at the database level. Make sure that:
Tables are structured properly, columns have the right data types.
Right indexes are in place to make queries efficient.
You are using the appropriate storage engine for each table.
You use an appropriate row format.
The application uses an appropriate locking strategy.
All memory areas are used for caching sized correctly.
Optimize at the hardware level. System bottlenecks typically arise from these sources:
Disk seeks. To optimize seek time, distribute the data onto more than one disk.
Disk reading and writing. When the disk is at the correct position, we need to read or write the data. With modern disks, one disk delivers at least 10–20MB/s throughput. This is easier to optimize than seeks because you can read in parallel from multiple disks.
CPU cycles. Having large tables compared to the amount of memory is the most common limiting factor. But with small tables, speed is usually not the problem.
Memory bandwidth. When the CPU needs more data than can fit in the CPU cache, the main memory bandwidth may become a bottleneck.
More recommendations are available in MySQL topic in Oro documentation.
For more information on performance optimization on MySQL website, see the Optimization section of the Reference Manual.
You can make Symfony faster if you optimize your servers and applications:
Use the OPcache byte code cache to avoid having to recompile PHP files for every request
Configure OPcache for maximum performance
1; php.ini 2; maximum memory that OPcache can use to store compiled PHP files 3opcache.memory_consumption=256 4 5; maximum number of files that can be stored in the cache 6opcache.max_accelerated_files=20000
Do not check PHP files timestamps. By default, OPcache checks if cached files have changed their contents since they were cached. This check introduces some overhead that can be avoided as follows:
1; php.ini 2opcache.validate_timestamps=0
After each deploy, empty and regenerate the cache of OPcache.
Configure the PHP realpath cache
1; php.ini 2; maximum memory allocated to store the results 3realpath_cache_size=4096K 4 5; save the results for 10 minutes (600 seconds) 6realpath_cache_ttl=600
Optimize Composer autoloader
1composer dump-autoload --optimize --no-dev --classmap-authoritative
For more information on Symfony performance optimization, see the list of all recommendations on Symfony website.
Improve Doctrine Performance¶
There are several things you can do to improve Doctrine performance:
Use the EXTRA_LAZY fetch-mode feature for collections to avoid performance and memory problems initializing references to large collections.
Mark a many-to-one or one-to-one association as fetched temporarily to batch fetch these entities using a WHERE ..IN query.
1<?php 2$query = $em->createQuery("SELECT u FROM MyProject\User u"); 3$query->setFetchMode("MyProject\User", "address", \Doctrine\ORM\Mapping\ClassMetadata::FETCH_EAGER); 4$query->execute();
More recommendations on improving Doctrine performance are available on Doctrine website.
Optimize Message Queue Consumers¶
MQ consumers may take up quite a lot of CPU time. To avoid this, consider moving consumers to a separate node, or have enough CPU cores in the main node.
Use Blackfire to Profile Requests¶
You can use Blackfire at any stage of application’s lifecycle to gather data about the behavior of your current codebase, analyze profiles and optimize the code.
Using Blackfire, you can find and fix performance issues by using the following methods:
Profile key pages
Select the slowest ones
Compare and analyze profiles to spot differences and bottlenecks (on all dimensions)
Find the biggest bottlenecks
Try to fix the issue or improve the overall performance
Check that tests are not broken
Generate a profile of the updated version of the code
Compare the new profile with the first one
Rinse and repeat
Read more on how to use Blackfire in its documentation portal.