• v2.3
    • Versions
    • master
    • v3.2
    • v3.1
    • v3.0
    • v2.3
    • v2.2
    • v2.1
    • v2.0
    • v1.7
    • v1.6
    • v1.5
    • v1.4
    • v1.3
    • v1.2
    • v1.1
    • v1.0

 

  • Install Akeneo PIM
    • Install Akeneo PIM with Docker
    • Install Akeneo PIM manually
      • System Requirements
      • System installation on Debian 9 (Stretch)
      • System installation on Ubuntu 16.04 (Xenial Xerus)
      • Installing Akeneo PIM Community Edition (CE) with the Archive
      • Installing Akeneo PIM Enterprise Edition (EE) with the Archive
      • Setting up the job queue daemon
    • How to customize the Dataset
    • How to Add Translation Packs
  • Migrate Akeneo PIM projects
    • How to apply a patch?
      • How to apply a patch - Community Edition
      • How to apply a patch - Enterprise Edition - Flexibility Cloud offer
      • How to apply a patch - Enterprise Edition - On Premise offer
    • Where is the Changelog?
    • How to upgrade to a minor version?
    • How to upgrade to a major version?
  • Import and Export data
    • How import works
    • Understanding the Product Import
    • Understanding the Product Export
    • Formats
      • Localized labels
      • Scopable labels
      • Association types data structure
      • Attribute data structure
      • Category data structure
      • Family data structure
      • Family variant data structure
      • Group data structure
      • Options data structure
      • Product data structure
      • Product model data structure
    • Akeneo Connectors
    • How to Customize Import / Export
      • How to create a new Connector
      • How to import Products from a XML file
      • How to clean a CSV file during a Product import
      • How to automate imports/exports
  • Manipulate the Akeneo PIM data
    • How to Customize Mass Edit Operations
      • How to register a new bulk action
      • How to Register a New Mass Edit Action on Products
      • How to Avoid Rules Execution on Mass Edit Actions
    • How to Manipulate Products
      • How to Query Products
      • How to Create Products
      • How to Update Products
      • How to Validate Products
      • How to Save Products
      • How to Remove Products
    • How to Manipulate Non-Product Objects
      • How to Query Non-Product Objects
      • How to Create Non-Product Objects
      • How to Update Non-Product Objects
      • How to Validate Non-Product Objects
      • How to Save Non-Product Objects
      • How to Remove Non-Product Objects
    • How to add a custom action rule
      • General information about rule format
      • How to add a custom action in the rule engine
    • How to Define Access Control List
    • How to Customize the Catalog Structure
      • How to Create a Reference Data
      • How to add a custom unit of measure
      • How to Add New Properties to a Category
    • How to Customize Product Assets
      • How to Add a New Transformation
      • How to Add a Default Thumbnail For Unknown File Types
      • How to connect to an external server for storage
      • How to change the validation rule to match a reference file to an asset
      • How to Mass Import Assets
    • How To Customize Teamwork Assistant (Enterprise Edition)
      • Customize notifications
      • Add a calculation step
      • How to log calculation step
      • Remove projects impacted by a custom catalog update
  • Design the user interfaces
    • How to customize any frontend part of the application
    • How to add an action button or meta data to the product edit form
    • How to add a tab to a form
    • How to add a new tab in System / Configuration
    • How to add custom information to a field
    • How to add a new field type
    • Create a custom product export builder filter
    • How to create the UI to manage a Reference Data
    • How to add a new page
    • How to customize the main menu
    • Styleguide
  • Maintain Akeneo PIM projects
    • First aid kit
    • Bug qualification
    • Common issues
    • Scalability Guide
      • Audit with 3 Representative Catalogs
      • More than 10k attributes?
      • More than 10k families?
      • More than 10k categories?
      • More than 500 attributes usable in the product grids?
      • More than 100k products to export?
      • More than 1GB of product media to export?
    • How to purge history
      • How to Purge jobs executions
      • How to adapt the version purger to your needs
  • Contribute to Akeneo PIM
    • How to report an issue?
    • How to translate the user interface?
    • How to enhance the documentation?
    • How to contribute to a Connector?
    • How to submit a patch to the PIM?
    • How to contribute to the frontend part of the application
    • How behavior tests are architectured in the PIM?
      • Establishing Decorator Pattern
      • Using Spin In Behat
  • Technical overview
    • Product Information
    • Teamwork Assistant (Enterprise Edition)
      • Project creation
      • Project completeness
      • Project Completeness widget
      • Catalog update impact
      • Scalability guide
      • Users permission summary for Behat tests
    • Collaborative workflow
      • Simple workflow
      • Partial workflow
  • Technical architecture
    • Best Practices
      • Create a project
      • Create a reusable bundle
      • Setup Behat
      • Code Conventions
      • Coding Standards
    • Bundles
    • Components
    • How to implement your business logic using the event system
    • Events
      • Storage events
      • Workflow events (Enterprise Edition only)
    • How to Localize your data
      • How to change the PIM locale
      • How to Use Localizers
      • How to use Presenters
    • How to Add a Notification
    • Performances Guide
      • Memory usage of rules execution (Enterprise Edition)
      • Memory leak fix in Rules Engine (ORM)
      • More than 100 WYSIWYG editors in a page
      • PHP7 and HHVM Compatibility?
      • Job product batch size
    • How to Use the Web REST API
    • Standard format
      • Products
      • Other entities
      • Usage
    • Application Technical Information
      • Application Technical Dependencies
      • Server side set up for hosting
      • System Requirements
      • Recommended configuration
      • Client side configuration and compatibilities
      • Operation processes
      • Flow Matrix
  • Akeneo Cloud Edition
    • Flexibility Mode
      • Overview
      • Partners Starterkit
      • Environments Access
      • Composer settings
      • Periodic tasks / Crontab settings
      • PIM Application
      • PIM Updates and Migrations
      • Data File Transfer
      • Backups management
      • Partners
      • Queue management
      • System Components
      • Disk Usage
    • Serenity Mode
      • Overview
      • PIM Updates and Migrations
  • Akeneo Onboarder
    • Prerequisites
    • Installation
    • Synchronization
    • Troubleshooting
    • Environment variables
      • Using the DotEnv file
      • Using environment variables

Audit with 3 Representative Catalogs¶

Warning

This page is an early version, we’ll continue to complete it with more use cases.

We’ve audited the application with 3 different representative catalogs:

Catalog Small Medium Large
Products 5.000 50.000 1.000.000
Categories 500 2.000 4.000
Categories / product 2 2 4
Attributes 100 400 1.000
Attribute Groups 8 15 20
Attributes / Families 50 100 100
% filled attributes 75% 75% 50%
%localisable attributes 10% 5% 2%
%scopable attributes 5% 2% 1%
%scopable + localisable attributes 2% 1% < 1%
Families 20 50 400
Channels 2 2 2
Enabled Locales 1 4 4
Audit Status for Community Edition Ok Ok Ok
Audit Status for Enterprise Edition Ok Ok Ok

These catalogs are available in our dedicated repository https://github.com/akeneo/catalogs, you can also use our data generator to build your own test catalog https://github.com/akeneo-labs/DataGeneratorBundle.

Several of our customers strongly push these limitations in their custom projects, you can consult different use cases in this page. We adapt these representative catalogs between minor versions when we improve the application scalability. Don’t hesitate to consult this page for other versions.

How we tested?¶

Installation

The application is installed on a server following the recommended architecture Application Technical Information.

Depending on the catalog, we installed the data fixtures via the installer before importing the products through the default product csv import job (for a large product import, we split the catalog into 10 files + parallel imports + custom optimisations).

  Product values
Small 159.676
Medium 3.661.981
Large 52.699.463

Audit User Interface

We use the application in production mode, with xdebug disabled, and we expect an optimal user experience for each page and action.

Audit Backend Processes

We run backend processes (bulk actions, imports, exports, rules execution, etc) in production mode, with xdebug disabled. Processes may run for quite a long time depending on the amount of data but do not consume more memory than the volume advised in Application Technical Information. Please note that for some project data set, several extra configurations are required (for instance, changing the size of a bulk of products for the rules execution).

Automation

We have built several tools to automate these performance and scalability tests. Basically, our continuous integration loads a target catalog and then runs different scenarios. The build is considered to fail when thresholds on time execution and memory usage are reached. These tools are not open sourced for now.

Known limitations on representative catalogs¶

We observed two kinds of limitations during the audit: memory leaks and scalability limitations.

Memory leaks when a process consumes more memory than recommended. These issues are qualified as bugs, their fixes are released in version patches.

Scalability limitations, when we try to support larger data volume for an axis. These issues are qualified as improvements which are released in minor versions.

The following limitations have been encountered with standard installations without any custom code:

Type Catalog Edition Released Note
memleak All All v1.4.23 (PIM-5507) Memory leak when executing mass edit or mass publish on several thousands of products
improv. Medium All TODO (PIM-5467) First load of the completeness widget is too long (ORM)
improv. Large All TODO (PIM-5518) Timeout with synchronous update of products when remove ‘AssociationType’, ‘Attribute’, ‘AttributeOption’, ‘Category’, ‘Family’, ‘Group’, ‘Channel’
improv. Large All v1.6.0 (PIM-5542) the request /configuration/family/rest slow down the UI (dashboard, grid, pef)
improv. Large Enterprise TODO (PIM-5544) the request /enrich/product-category-tree/list-tree.json allowing to load the tree on the grid is very slow (improved with Elastic Search Bundle)
improv. Large All TODO MongoDB timeout when filtering and sorting on product grid when using not indexed fields (improved with Elastic Search Bundle)

Examples of customers instance¶

Several customers challenge the limitations even more in their custom projects and it requires custom optimizations sometimes. We continuously improve the product scalability in each minor version and we are always interested in new use cases to investigate. Don’t hesitate to contact us if you need help to scale your instance.

On standard axes:

Catalog Customer 1 Customer 2 Customer 3 Customer 4 Details about limitations
Storage MongoDB + ES MongoDB + ES MySQL MySQL ES: ElasticSearch Bundle
Products 2.000.000 1.100.041 80.000 10.000  
Products values 43.398.847 78.606.501 6.000.000 70.000 6 millions product values is a high limit for MySQL storage
Attributes 1.800 8.272 240 355 More than 10k attributes?
Families 131 3.546 44 3 More than 10k families?
Categories 2613 14.238 740 60 More than 10k categories?
Channels 1 2 2 14  
Enabled Locales 1 1 36 1  

On other axes or combinations:

  Tested In custom project Details about limitations
Attribute options 95.000    
Attribute options per attribute 500    
Reference data [WIP]    
Reference data per attribute [WIP]    
Products per family [WIP] 1.000.000 cf following PIM-5563
Product groups 10.000   cf following PIM-5519, PIM-5363
Products per product group 50    
Product variant groups 10.000   cf following PIM-5467, PIM-5520, PIM-5363
Products per product variant group 50    
Product values per variant group 50    
Product associations [WIP]   cf following PIM-5363, PIM-5562
Attributes per family 150    
Attributes per attribute group 150 1.500  
Product values per product 200    
Rules 150 3.000 Memory usage of rules execution (Enterprise Edition)
Product assets [WIP]    
Product drafts [WIP]    

Known limitations on other axes or combinations

Type Catalog Edition Released Note
improv. All All TODO (PIM-5519) Mass edit products, display the add to a group configuration is too long with a lot of product groups (use a paginated select2 and not checkboxes)
improv. All All TODO (PIM-5520) Mass edit products, display the add to a variant group configuration is too long with a lot of product groups (use a paginated select2)
improv. All All TODO (PIM-5467) When saving a variant group, variant group values are synchronously copied in products, it may cause timeout issue
improv. All All TODO (PIM-5463) When associating a lot of products to a group, variant group or association, you may encounter “The requested URL’s length exceeds the capacity”
improv. All All TODO (PIM-5562) When delete a product associated to other products, run a backend process to cleanup all associations
improv. All All TODO (PIM-5563) Query for completeness rescheduling when saving a family with 50k products inside is too long to execute, exec as backend process
improv. All All TODO (PIM-5861) Cannot remove an attribute on big (~1M products) MongoDB catalog because of timeout when we update MongoDB database to remove it from all products
improv. All All TODO (IM-766) Variant groups scalability: limit number of axes or set a limit?

Found a typo or a hole in the documentation and feel like contributing?
Join us on Github!