OVERVIEW
With the onset of COVID-19, traditional sales channels witnessed a drop, and the focus shifted on providing an exceptional online customer experience, with organizations observing a 46% YoY growth in online orders. Add to this, the retailers moved from an on-premise deployment to Google Cloud Platform (GCP) and had to make sure that the shift was in line with their scalability needs in terms of handling real-world user traffic.
Furthermore, projections at retailers estimated that compared with 2019, the year 2020 would see a huge increase in revenue generation via online channels (~6x times), and the organization could not afford downtime to affect customer experience and, in turn, revenue.
PERFORMANCE TESTING PLATFORM TO HANDLE UNPRECEDENTED LOAD
One of the largest retailers in the US, it excels in customer experience by using Cavisson’s performance testing platform during their peak season. We were instrumental in mimicking real-world behavior and ensuring that the scalability and resilience of the system remain intact.
A holistic approach adopted to test the entire retail ecosystem comprising of in- store and online channels, generating 100+ million hits/hour, thereby delivering exceptional CX.
SOLUTION
Cavisson’s team of performance experts coordinated with retailer’s performance team to mimic real-world traffic in their pre-production environment to determine the readiness of the application.
A detailed step-by-step playbook was prepared to handle negative scenarios. All previously used scripts were modified to reflect the changing scenario.
The CI/CD-enabled Performance Test Lab was established to reduce time & cost. This pipeline was instrumental in achieving the following:
Generated 108 million transactions/hour and 36 million orders/hour via Cavisson testing scripts.
Individual domain scripts were created for capacity tuning across different channels (desktop, mobile, and tablets).
Underlying infrastructure was scaled up & down automatically depending upon the demand/load.
Analyzed trends & mixes from previous years’ data, including orders/min, TPS, cart size, and traffic patterns, to replicate real-world traffic.
Automated reports published on Jenkins & sent to users post-test
Monitored peak period and prepared recommendations for the enterprise’s application development team to further fine-tune their application and infrastructure for future readiness.
CUSTOMER INFO
Industry: Retail
Location: United States
Challenge: Ensure systems readiness for high sales days, where OPM (order per minute) reaches 200 times in comparison to normal sale days.
Product Used: Cavisson Performance Testing
Result: Witnessed zero downtime by adopting a concerted strategy to tackle real-world traffic and giving clear indications of system readiness.
BENEFITS
Confident in its load testing capabilities, retailers witnessed a successful Cyber CyberMonday and Thanksgiving wherein ZERO downtime was observed, thereby resulting in increased revenue.
NetOcean’s API capabilities were instrumental in testing POS systems deployed at stores to check their readiness in terms of handling the increased load expected during the peak season by testing APIs that interacted with warehouses, stores, and retailer’s e-commerce platforms.
For detailed analysis, Java Flight Recorder and thread and heap dumps were taken in order to reduce MTTD and MTTR in retailer pre-production and production environments.
Performance and exceptions across different domains were analyzed, including application thread pools, connection pools, memory & system usage, along with network analysis and database performance, to give the retailer a 360-degree view of their applications during the peak season.
