We can see that a large enterprise customer implemented a MongoDB sharded cluster on SolidFire as the backend for a global e-commerce system, and they were experiencing response times of around 15-20ms, which was deemed out of range for their application's performance requirements. To troubleshoot this issue, the customer reached out to SolidFire and MongoDB for assistance with configuration. After analyzing the end-to-end chain, including potential i/o queue depth bottlenecks and the main contributors to overall latency, they discovered that the primary cause of the high latency was the MongoDB flush interval, which defaults to 60 seconds. By increasing this interval from 60 seconds to just under 1 second, the customer was able to reduce their average end-to-end MongoDB response times by more than an order of magnitude, and improve their throughput and IOPS per mongod. This change had a significant impact on reducing latency, with some segments experiencing a decrease of over 3x, while others saw a reduction of around 3.5-4x. The customer's experience demonstrates the importance of early engagement with MongoDB and highlights SolidFire as a MongoDB partner.