Case Study: PhonePe achieves 10x open-source data platform scale with Acceldata

Phonepe Uses Acceldata to Scale Open-source Data Platform by 10x

PhonePe (Walmart), which handles roughly 400M cash transactions per month on a large open-source data stack (Hadoop, HBase, Hive, Spark, Ranger, Kafka) and an HBase OLTP cluster at ~1.1K TPS, faced scaling and performance issues across OLTP and OLAP environments. To diagnose and distinguish infrastructure problems from seasonal and campaign-driven anomalies, PhonePe engaged Acceldata and deployed the Acceldata Pulse data observability solution.

Using Acceldata Pulse to monitor HBase, Spark, and Kafka, Acceldata helped PhonePe scale its data infrastructure 10x (from ~70 to ~700 nodes), achieve 99.97% availability, save more than $5M annually in software licensing, and reduce daily firefighting to boost engineering productivity—enabling massive business expansion of one of the world’s largest instant payment systems.

PhonePe

Burzin Engineer

Co-founder and Chief Reliability Officer,