Case Study: How We Achieved 52% Magento Hosting Cost Reduction

Many e-commerce store owners believe reducing Magento hosting cost reduction requires server upgrades, but scaling up often inflates monthly bills without addressing core performance issues.

In this case study, Growsera’s Magento support team improved a client’s Magento store performance through Magento Elasticsearch tuning and MySQL/MariaDB optimization, cutting server resource usage and achieving Magento hosting cost reduction of over 52%.

Challenges We Faced

1. Frequent Elasticsearch Crashes

Elasticsearch crashed every few hours, even with low traffic, disrupting search, hiding products, and impacting Magento admin panel operations.

2. Resource Overload

The store consumed 3.9GB of 4GB RAM, with ~85% CPU usage, causing unstable hosting and frequent 524 and 500 internal server errors.

3. Misconfigured Services

The previous technical team installed Elasticsearch, Redis, and MariaDB without proper configurations for Magento 2, leading to performance bottlenecks.

4. Costly, Ineffective Proposal

The prior team suggested upgrading to a t3.large or t3.xlarge server, doubling monthly costs without addressing root issues.

Root Solution for Magento Hosting Cost Reduction

1. Optimized Elasticsearch with 50% Heap Rule + Module Cleanup
   • Reduced Xms/Xmx to 1.8GB (from 4GB), aligning with Elasticsearch’s 50% Heap recommendation.
   • Removed unnecessary Magento modules (e.g., ingest-geoip, ingest-user-agent) using elasticsearch-plugin remove <plugin>, cutting JVM footprint by ~250MB and reducing compilation time.
   • Set indices.query.bool.max_clause_count to limit costly queries and closed unused indexes, enhancing Magento Elasticsearch.
2. Proper MySQL/MariaDB Configuration
   • Set innodb_buffer_pool_size to 2.2GB ≈ (55% of RAM), per Percona Community Forum guidelines for 50-70% of available memory.
   • Disabled query_cache to reduce memory usage and enabled slow_query_log with a 2-second threshold to optimize SQL queries, per MySQL documentation.
   • Optimized queries in modules like Webkul_Marketplace for better performance optimization.
3. Additional Improvements
   • Disabled Transparent Huge Pages and set vm.swappiness to 10 to minimize swapping.
   • Integrated New Relic for real-time alerts when server resources exceed 70%, supporting Magento maintenance.
4. Flexible Action Plan (Quick Fix + Flex Hours)
   • Fixed urgent crashes and tuned JVM within 24 hours via Quick Fix tickets.
   • Used 8 Flex Hours to review indexes and optimize the database, ensuring Magento hosting cost reduction.

These solutions achieved Magento hosting cost reduction of 52% on the client’s monthly bill.

Results by the Numbers

Metric	Before Tuning	After Tuning
Server Size	t3.large	t3.medium
Average CPU (Peak)	85%	41%
RAM Usage	3.9GB	1.8GB
AWS Hourly Cost	$0.0832	$0.0416
Weekly Elasticsearch Crashes	14	0

Final Outcome: Over 52% Magento hosting cost reduction, improved store stability, and faster performance.

Business Benefits for You

• Immediate Savings: Halving resource usage cuts server costs, avoiding unnecessary upgrades with performance optimization.
• Higher Stability, More Sales: Zero downtime ensures no lost customers, supported by Magento maintenance.
• Seasonal Scalability: With Infrastructure as Code, resources can scale temporarily for peak seasons, reverting automatically to save costs.

Actionable Steps for Your Store

1. Set Heap to ≤50% of RAM and ensure Xms = Xmx.
2. Remove unused Elasticsearch modules before considering server upgrades.
3. Configure innodb_buffer_pool_size to 50-60% of RAM, reserving the rest for File System Cache.
4. Monitor Slow Queries regularly, especially after adding new Magento modules, for ongoing maintenance.

FAQs

Ready for a stable, cost-efficient Magento store? Try Growsera’s exceptional Magento support services. [Explore Our Support Plans]