Case Study: Razorpay achieves real-time payment analytics and lower infrastructure costs with StarTree Cloud

Razorpay cuts AWS upsert infrastructure costs by 50% with StarTree

Razorpay, India's fastest-growing payment processing company, needed to monitor payment success rates and analyze financial transactions in real time across over 100 different payment methods. Their previous solutions, based on Elasticsearch and Postgres, only provided data that was 15 to 20 minutes old. To achieve true real-time analytics, they turned to vendor StarTree and implemented StarTree Cloud, powered by Apache Pinot, and StarTree ThirdEye for anomaly detection.

StarTree implemented a real-time analytics architecture for Razorpay that ingests 200 million events daily from various databases via Kafka. A key feature was leveraging StarTree Cloud's off-heap upserts, which handle updates and inserts scalably without the memory pressure of open-source alternatives. This solution provided Razorpay with sub-second query latencies for user-facing dashboards and enabled proactive anomaly detection. As a result, StarTree helped Razorpay reduce its infrastructure costs for upsert tables by 50%.

Razorpay