5 Hard-Learned Lessons from Successful Database Migration and Modernization
summary
Singtel migrates production identity and LDAP data to YugabyteDB using a Voer migration pipeline and automated Puppet pipelines, EFS+S3 backups, Prometheus/Grafana and ELK for observability. A router service keeps dual writes during cutover and steers reads by percentage to avoid disruption. Stress tests hit 4,000 TPS on single-table reads, 1.3k TPS on 20-table joins, and 13.5k TPS writes (15 min until disk full); combined load sustains ~4,000 TPS at ~266 ms. Pain points: backup gaps, missing metrics, and schema/type mismatches requiring custom functions.
AI-written summary · about
chapters (14)
- 0:00Database migration journey
- 1:35YugaByte DB proof of concept
- 5:10Choosing cluster setup
- 9:47Production cluster setup
- 10:51Automating node provisioning
- 13:00Custom observability tools
- 14:29Legacy LDAP to YugaByte
- 16:37Migration strategy with Router
- 20:00Yugabyte Voyager usage
- 25:22Data consistency post-migration
- 25:42Yugabyte Console integration
- 26:25Stress testing YugaByte
- 33:28YugaByte challenges learned
- 35:39Conclusion and benefits
related talks
22mMigrating from Monolithic to Cloud Native Distributed SQL Database with Zero Downtime
31mMake it Painless: YugabyteDB Database Migration Solutions
35mA Customer’s Journey: Migrating from NoSQL to a Cloud Native Distributed SQL Database
27mSparking our way to YugabyteDB: Our journey from ScyllaDB to
35mScaling UPI for the Future: Zero Downtime Migration to
13mStrategic Shift to YugabyteDB as Core Database
19mEvolve: A Database Journey from Ground to Cloud
31mYugabyteDB Voyager: Simplify Data Migration to Accelerate Cloud Adoption