Case Study: Apna achieves near real-time, more reliable analytics with Onehouse

Apna delivers near real-time data with Onehouse for 33 million users

Apna, a major professional networking and job matching platform in India, faced significant challenges with its legacy data architecture built on Google BigQuery. The batch-oriented system resulted in poor data freshness, high costs for near real-time synchronization, complex ETL failures, and limited time travel capabilities. These issues hindered the company's ability to provide timely data for its AI-powered job matching engine and other critical data products, creating multiple points of failure and operational complexity.

To overcome these challenges, Apna implemented the Onehouse managed service to build a new Universal Data Lakehouse platform called Apna Analytics Ground (AAG). Onehouse provided managed ingestion from Postgres, MongoDB, and Kafka, automatically streaming data into managed Apache Hudi tables on Google Cloud Storage. The solution delivered data in near real-time with tunable SLAs, eliminated multiple failure points, reduced costs through decoupled storage and compute, and provided long-term time travel capabilities. The new architecture, powered by Onehouse, made data more consistent and available while allowing users to continue leveraging BigQuery and easily add new compute engines like Spark and Presto.

Apna