Case Study: Meesho achieves sub-second real-time analytics with StarTree and Apache Pinot

Meesho achieves sub-second real-time analytics with StarTree

The customer Meesho, a major e-commerce platform, faced a challenge as their initial analytics system using Apache Druid and Flink was limited to pre-aggregated data. This prevented granular, root-cause analysis of user behavior, creating bottlenecks for their growing need for real-time insights. They sought a new solution to handle raw data exploration at scale.

The vendor StarTree, through an implementation of Apache Pinot, provided the solution. This enabled Meesho to build a new real-time analytics platform featuring a denormalized data model and a custom federation layer for query routing. The results included achieving sub-second P90 query latencies, which empowered teams with faster incident management and granular root-cause analysis on massive volumes of event data.

Meesho