blue-star-01-1

Extreme scale engineering

Discover the latest trends and best practices impacting data-intensive applications. Register for access to all 60+ sessions available on demand.

From Batch to Real-Time: How MoEngage Achieved Millisecond Personalization with ScyllaDB

Atish Andhare14 minutes
Share this
Share this

Register for access to all 60+ sessions available on demand.

Fill out the form to watch this session from the Monster Scale Summit livestream. You’ll also get access to all available recordings.

In this Monster Scale Summit Presentation

At MoEngage, the delay between a user action and our ability to react to it is the difference between a conversion and a missed opportunity. Our mission to power real-time customer engagement was hindered by a slow, batch-oriented data pipeline that couldn't keep up with our massive scale. To solve this, we built our "Eventstore" platform on ScyllaDB to handle 250,000 writes per second at 1ms p99 latency to store over 200TB of raw data, architecting a purpose-built schema to tame the challenge of multi-tenancy. This new system guarantees zero data loss through a cross-AZ replication strategy and simplifies cluster management at scale by leveraging ScyllaDB's incremental compaction strategy.

Atish Andhare, Technical Lead, MoEngage

Atish is technical lead at MoEngage, focusing on online datastores