Last updated:
Benum: Magento Performance Optimization with 6x Faster Page Loads and 94% Lower Database Response Time
Adobe Commerce (Magento) performance optimization for a Nordic B2B distributor of professional audio equipment: database query rewrites, a separate database on Amazon RDS for MariaDB, Varnish and Redis caching, AWS capacity tuning and load testing.
Project Summary
Elogic Commerce optimized the performance of Benum’s Adobe Commerce store, a large B2B site with more than 1.6 million product URLs and over 120 customer groups. As Benum added new technologies and custom modules over the years, some pages took as long as 25 seconds to load, contributing to lost prospects and sales.
Elogic Commerce found three causes: database queries that became inefficient after a platform upgrade, a production server that was running out of resources at peak times, and caching that struggled at the store’s scale. Rewriting the queries and moving the database to a separate Amazon RDS for MariaDB instance, alongside Varnish and Redis caching, made average page load time across the site six times faster and cut database response time from about 5 seconds to 0.3 seconds. Optimization continues, with Elasticsearch moving to its own instance as the next step.
Key Outcomes
6x
faster average page load time across the site after optimization
−94%
database response time, from about 5 seconds to 0.3 seconds
−40%
peak application-server resource utilization (CPU, RAM and load) after database migration and infrastructure optimization
1.6M+ and 120+
product URLs and customer groups on one optimized Adobe Commerce store
About the Client
Based in Oslo, Norway, Benum is a leading distributor of professional audio equipment in the Nordics. It started online with a small business solution, then migrated to Adobe Commerce in 2020 with Elogic. Since then, Elogic has redesigned its UI/UX around the buyer’s journey, integrated Visma Business ERP and built custom modules.
See the main Benum case study for the migration and ERP integration.
Project Complexity
Pages could take up to 25 seconds to load, leading to lost prospects and low sales.
As modules and integrations were added over the years, the site slowed down gradually, and no single change was the cause.
With 1.6 million product URLs, 120+ customer groups and shared catalogs, standard caching and tuning approaches were difficult to scale.
After platform upgrades, some modules were behaving incorrectly with the database.
All changes had to be made on a live B2B store used daily by business buyers.
Business Challenge
01
Page loading had degraded to as much as 25 seconds, putting online sales at risk
02
The root causes were spread across database queries, caching and AWS infrastructure
03
The store's scale made conventional caching and performance tuning increasingly difficult to maintain
04
The database shared a production server that had only 5% free capacity at peak times
05
Every fix had to be delivered without disrupting daily B2B operations
Elogic Commerce's Solution
Rewriting database queries
After a platform upgrade, some modules were behaving incorrectly and needed extra query optimization. At Benum, the Mageplaza BetterSorting module and Magento’s Multi-Source Inventory (MSI) were slowing down responses, so Elogic rewrote their database queries for the new platform version.
Server resources and load testing
Benum’s store was hosted on AWS, but its production server had only 5% free capacity during peak periods. Elogic added AWS server resources and recommended regular load and stress testing to prevent future bottlenecks.
Rethinking caching
With about 1.6 million product URLs across more than 120 customer groups and shared catalogs, cache configuration and warm-up were particularly challenging, and caching could never expose one buyer’s prices to another.
Elogic Commerce optimized the Varnish caching strategy to support Benum’s large B2B catalog and 120+ customer groups without compromising customer-specific pricing or catalog visibility. Common product content can be reused across requests, while personalized pricing and other customer-specific information stay tied to the right buyer context.
Moving the database to MariaDB on Amazon RDS
Benum’s MySQL database originally ran on the same production server as the Adobe Commerce storefront, adding resource pressure as the platform grew, and database requests took about five seconds on average. Elogic separated the database from the application server, migrated it from MySQL to MariaDB and deployed it on Amazon RDS for MariaDB, which also simplified ongoing database administration through managed database capabilities.
What’s next
Performance optimization is an ongoing process that needs constant monitoring. The next step is moving Elasticsearch to a separate instance.
| Bottleneck | Fix |
|---|---|
| Database queries slowed by BetterSorting and MSI | Queries rewritten for the new platform version |
| Production server with 5% free capacity at peak | Additional AWS resources and regular load and stress tests |
| Caching at 1.6M product URLs and 120+ customer groups | Varnish caching strategy that reuses common content and keeps customer-specific prices per buyer |
| MySQL on the same server as the storefront | Separate database on Amazon RDS for MariaDB, with Redis caching |
Results & Business Impact
Faster page loading
Following the database migration and performance optimization, average page load time across the site became six times faster. Before the work, the slowest pages took up to 25 seconds to load.
Faster database responses
Moving the database to Amazon RDS for MariaDB, alongside query optimization and Redis caching, cut database response time from about 5 seconds to 0.3 seconds.
A more scalable infrastructure
With the database on its own managed instance, resource contention dropped and the infrastructure became easier to manage for Benum's large B2B catalog.
A foundation for further optimization
The work removed critical database and infrastructure bottlenecks and set up the next steps, starting with moving Elasticsearch to its own instance.
Lower server resource usage
Peak application-server resource utilization, covering CPU, RAM and load, fell by 40% after separating the database, adding AWS resources, optimizing queries and improving caching, supporting more stable performance during high-traffic periods.
Capabilities Demonstrated
01
Magento and Adobe Commerce performance audit
02
Database query optimization after platform upgrades
03
Database migration from MySQL to Amazon RDS for MariaDB
04
Varnish and Redis caching strategy for very large B2B catalogs with customer-specific pricing
05
Load and stress testing on AWS
06
Ongoing performance monitoring
When This Solution Is a Good Fit
Performance optimization like this works best when the store’s design and business logic are sound but speed has degraded as the platform, modules and traffic have grown. The audit should come first, so the fixes target real bottlenecks, and monitoring should continue afterwards.
It is generally not the right fit when the platform itself no longer matches the business, where a replatforming project will deliver more than further tuning.
Planning an Ecommerce Transformation?
Elogic Commerce can audit your Adobe Commerce store, find the real bottlenecks and fix them without disrupting your business. Contact us to discuss your project.