Why Squarespace adopted CockroachDB for better global performance at scale
summary
Squarespace migrates ~50 Postgres clusters with about 100 individual cutovers using Debezium CDC into CockroachDB and reports 80% of clusters require minimal app changes. The pipeline copies snapshots, streams WAL into Kafka, transforms and writes into Cockroach, then switches clients via a managed proxy (HAProxy) while validating data and enabling a rollback feed. Outcomes: online binary and schema upgrades, daily console visibility, fewer incidents, and multi‑region failover without lost writes.
AI-written summary · about
chapters (16)
- 0:00Squarespace and talk intro
- 0:31Breaking up with PostgreSQL
- 1:56Vertical scaling challenges
- 3:00Global topology benefits
- 4:23PostgreSQL operational issues
- 7:52CockroachDB vs managed PostgreSQL
- 8:57CockroachDB’s advantages for application teams
- 10:45Online upgrades impact
- 11:06Migration overview
- 13:36Syncing and validation
- 15:01Client cutover testing
- 16:44Final cutover rollback
- 18:07Lessons learned buy-in
- 19:54Legacy patterns handling
- 20:16Developer tooling benefits
- 21:19Global gains summary
related talks
48mBuilding mission-critical applications with CockroachDB
27mBeyond the Roadmap and Into the Future of CockroachDB
28mGlobal Payments: Subsecond Performance Worldwide
18mFrom Today to 10X: Why Groww chose
24mNavigating Distributed Databases: Challenges & Opportunities
29mHow City Storage Systems designed its Data Infrastructure to minimize toil
17mMigrating Primer’s Live Payments Platform to CockroachDB with Zero Downtime
22mBuilding CockroachDB for performance at scale