Why Squarespace migrated global data workloads from PostgreSQL to CockroachDB

Every year at RoachFest, we hear directly from the teams running some of the most demanding data workloads in production. RoachFest 2025 was no different—except this year, a clear theme emerged across customer sessions: primary data and metadata are often mission-critical.

From global education platforms to internet-scale website builders, these companies are managing massive volumes of user-facing data that must be correct, highly available, and fast everywhere. Downtime is not an option. Neither is data inconsistency. And as these workloads grow, traditional database architectures start to show their limits.

At RoachFest 2025, customers shared how they’re using CockroachDB to meet these challenges head-on—migrating critical user data with zero downtime, simplifying global data infrastructure, and operating at scale without adding operational risk. 

Gavin McQuillan, Senior Staff Software Engineer at Squarespace, highlighted a story about how and why they migrated from PostgreSQL to CockroachDB, and why more internal teams are choosing CockroachDB as the foundation for their most important user-facing systems.

“Why Squarespace adopted CockroachDB for better global performance at scale”

Squarespace powers millions of websites worldwide, and much of that experience depends on globally distributed data, site configuration, user settings, and operational state that must be consistent, fast, and always available. As those workloads grew, PostgreSQL’s single-primary model became a limiting factor.

The team faced three core challenges:

• Vertical scaling limits, needed a strategy for horizontal scalability

• Scaling across different regions and continents for global audience

• Improved resilience, optimal RTO

• Operational complexity, particularly when upgrading with extensive downtime

• Lack of visibility via native observability tools

Image Source: Squarespace RoachFest 2025 Presentation based on their views of CockroachDB vs. Managed SQL's features

Managing failovers, upgrades, and backups across dozens of PostgreSQL clusters required significant manual effort and coordination, often with unavoidable downtime. For a platform with global customers, that risk was increasingly unacceptable.

CockroachDB offered a different model. With built-in replication, automated failover, and horizontal scale, Squarespace could treat primary data and metadata as globally available infrastructure rather than region-bound state. Online upgrades and schema changes eliminated planned downtime, while the CockroachDB console gave engineers visibility into how specific queries and services behaved across the system.

Migrating 50 staging clusters, then migrating those same 50 clusters in production, the Squarespace team has managed over 100 total migrations into CockroachDB. No queries dropped, no errors. Squarespace used change data capture, validation, and proxy-based cutovers to move safely without long outages. Most migrations required minimal application changes.

For Squarespace, CockroachDB now underpins critical metadata services with stronger resilience, simpler operations, and a clear path to global scale. The result is a database that fades into the background, while the platform continues to grow everywhere their customers are.

The key thread

Data demands correctness, availability, and scale at the same time. CockroachDB gives teams a way to meet those requirements without bolting together brittle workarounds or asking users to tolerate downtime while systems evolve.

If you are running user-facing workloads where data correctness and uptime matter, these kinds of stories show what it looks like to modernize without disruption and why more teams are choosing CockroachDB as their system of record.