Shopify: One of the World's Largest Distributed Postgres Databases

How YugabyteDB and Shopify are Pushing Distributed SQL Into New Territory Together

Distributed usually means scale, resilience, and multi-region deployments, fanning reads out across replicas while a single writer handles the rest and write-heavy workloads get pushed off to another system.

Shopify needed more: a single global namespace with multi-regional, quorum-based write transactions and no single point of failure.

Flexible placement strategies had to manage latency across regions while keeping global reads fast, all without forcing a rewrite of the application's relational modeling, a hard constraint after two decades on a single custom-sharded MySQL footprint.

Addressing these business needs meant pushing distributed SQL into new territory. A single Postgres-compatible database (YugabyteDB) now:

• Spans seven regions, distributing both reads and writes across roughly 180 production nodes and over 7,500 vCPUs, with a path to 20,000.

• Serves 600,000 queries per second with peaks near a million, at 1GB/sec throughput, with 42,400 tablet leaders across 250,000 tablet replicas, most 100GB or larger.

• Runs replication factors from 3 to 7, depending on the table

By total capacity in a single logical database, it is one of the largest known distributed Postgres deployments in production.

Why watch?

Brad Dietrich, Distinguished Engineer at Shopify, and Karthik Ranganathan, Co-founder and Co-CEO at Yugabyte, walk through the architectures both teams built and explore what agentic commerce demands of the data layer.

Save your seat by filling out the form, and join us at 11:00 am PT on Oct 14, 2026, to go inside the Shopify and YugabyteDB architecture.