How Helpshift Accelerated Growth by Migrating to
summary
Helpshift migrates to YugabyteDB and builds an 18-node cluster (≈4 TB) using Odyssey for connection pooling and HProxy for routing; sharding removes manual scaling, ACID transactions and per-query tunable consistency replace previous integrity problems, and JSONB handles semi-structured data. CDC runs via Kafka Connect into S3 per-table topics for eventual consistency with search, and write latency falls from seconds to roughly 30–40 ms in production while supporting ~82M active users and planning for 1B.
AI-written summary · about
chapters (10)
- 0:00Hail Shift introduction
- 1:23Legacy non-SQL challenges
- 4:16Database stack performance
- 8:18New database features
- 11:09Migration approach & schema
- 13:36YugabyteDB architecture
- 16:54Cluster optimizations
- 20:29ODC configuration
- 24:29Change Data Capture to ensure consistency
- 27:27Results & future outlook
related talks
27mSparking our way to YugabyteDB: Our journey from ScyllaDB to
29mNoSQL to Distributed SQL: Best Practices and Lessons Learned by Helpshift
35mA Customer’s Journey: Migrating from NoSQL to a Cloud Native Distributed SQL Database
22mMigrating from Monolithic to Cloud Native Distributed SQL Database with Zero Downtime
28mAccelerate and Secure Your Distributed Data Ecosystem with Red Hat OpenShift
19mEvolve: A Database Journey from Ground to Cloud
25mIncrease App Modernization Success with Red Hat Open Shift and YugabyteDB
20mDatabase migration: Insights on Dual Read Write Strategy